Ford rehires engineers it replaced with AI, Should we ban AI superintelligence?, AI reveals ancient manuscript’s secrets, How to use ChatGPT to avoid heated arguments, ChatGPT is American’s favorite AI tool, Two-thirds of workers are using banned AI tools in secret, Anthropic’s latest models released by Feds, AI glasses aid cheating on exams, mass AI fraud at Brown university, AI helps FBI investigate Correspondents’ Dinner shooting, and more, so…
These are the five prompts that fit any circumstance whenever I go to ChatGPT to make a final decision on buying something expensive. Hopefully, these commands do the same for you.
I’m considering buying [product name]. Help me identify the underlying problem I’m trying to solve and whether there are cheaper or simpler solutions.
Act as a skeptical consumer advocate. Give me the strongest reasons not to buy this product.
Describe the ideal buyer for this product and tell me whether I fit that profile.
Create a decision framework for this purchase. What factors should I compare and how should I weigh them?
Imagine it’s one year from now. What are the most likely reasons I would regret buying this product and the most likely reasons I’d be glad I bought it?
One of the easiest things to do during an argument is react. And in the heat of the moment, one of the hardest things to do is to pause.
Whether it’s a disagreement with your partner, a tense conversation with a coworker or a frustrating text message from a friend, emotions have a way of convincing us that the first thing we want to say is also the best thing to say.
I’ve learned that’s rarely true.
Instead of asking ChatGPT to tell me who’s right, I recently started using what I call the Pause Prompt. No, the goal here is not to win the argument, but rather to slow down just enough to response with intention instead of emotion.
After trying it several times, I can see myself coming back to it whenever emotions start taking over.
Here's the prompt I give ChatGPT: "I'm upset and I want to respond, but I don't want to make things worse. Here's what happened: [Describe the situation.]
Before giving advice:
• Tell me what emotions you think I'm experiencing. • Explain what the other person might be feeling, even if they expressed it poorly. • Point out any assumptions I'm making or information I may be missing. • Help me write a response that is honest, respectful and likely to move the conversation forward instead of escalating it.
Don't automatically agree with me. If I'm being unfair or reacting emotionally, tell me kindly but directly."
Most AI prompts ask ChatGPT to solve a problem. But this prompt is for you. It works because it helps you slow down.
OpenAI says ChatGPT has become less literal and more conversational after quietly updating GPT-5.5 Instant for everyone. Naturally, I wanted to know whether I could actually tell the difference.
After spending the day chatting with the updated model, both by text and in Voice Mode, I came away with mixed feelings. Some conversations did feel more natural, but others reminded me we’re still a long way from talking to AI as effortlessly as we talk to another person.
Rather than making ChatGPT dramatically smarter, OpenAI says the update should make it feel easier to talk to. The AI is supposed to infer what you’re trying to accomplish instead of taking every prompt literally, adapt more naturally when you change your mind, and keep track of the thread of a conversation without needing as many reminders.
Now, a lot of that will naturally fade into the background as you talk to ChatGPT, since none of it is designed to stand out, so it can be hard to quantify if the update is objectively better. In my own testing, I noticed ChatGPT felt quicker to adapt when I clarified what I wanted. Instead of sticking rigidly to its original interpretation, as it often does, it seemed more willing to adjust course as the conversation evolved.
I still have to tell it if it’s off the mark, but I get the feeling that it’s actually listening to me and is going to remember what I want and deliver it next time.
ChatGPT knows a lot about Connor Effrain. The bot is privy to every bathroom trip he takes, every morsel of food he consumes.
Effrain has Crohn’s disease, a chronic inflammatory bowel condition. His gastrointestinal tract mostly lays low, but he gets flare-ups every few months. The 22-year-old digital fundraising associate keeps ChatGPT apprised of his symptoms. Almost nothing remains secret between them.
The same could be said of his sophomore-year study partner: He shared his account for their academics, but soon realized she could also see his personal logs.
“I was like, ‘You know, I don’t like you knowing that I’m constipated in the morning,’” Effrain says he told her. He admits he never went over the ground rules before sharing his password, and he hasn’t shared his account since.
“I was a bit stupider back then,” he says.
It turns out, chatbot passwords deserve a bit more protection than a Netflix login.
AI companies, including ChatGPT maker OpenAI, frown on chatbot sharing. That hasn’t stopped college students cramming for exams together and couples collaborating on work projects from dishing out their passwords.
Yet embarrassing discoveries, personal data mix-ups and even cybersecurity issues abound, so you should understand the risks—and ways to stay safe.
The Trump administration has lifted restrictions on artificial intelligence company Anthropic’s latest versions of its Claude chatbot, ending a weekslong ban tied to cybersecurity concerns.
Anthropic said Tuesday night that its AI model called Claude Fable 5 is now widely available. It’s also restoring access to its most powerful model, Mythos 5, but only to a select group of U.S.-based organizations approved by the federal government.
The Commerce Department blocked foreign nationals from using both AI models on June 12, a move that San Francisco-based Anthropic said forced the company to immediately take the products down for all users just days after it unveiled them.
Anthropic said in a blog post this week that the government’s concerns were sparked by a report from cybersecurity researchers at Amazon, Anthropic’s primary cloud computing provider. The company “had found a method of bypassing Fable 5’s safeguards” that enabled it to discover and potentially exploit software vulnerabilities, Anthropic said.
Officials have grown increasingly concerned since Anthropic warned earlier this year that its Mythos model was adept at finding software flaws in a way that could be weaponized by malicious hackers and threaten critical computer networks around the world.
Americans are using chatbots more than ever before and some are bringing smart devices into their households, according to a new Pew Research Center survey of U.S. adults.
The key takeaways:
About half of U.S. adults now report using AI chatbots, up substantially from the summer of 2024.1 This includes roughly one-in-four who use these tools on daily basis.
Some people are bringing AI into their homes. About a third of Americans say they have a smart speaker, and smaller shares have a doorbell or thermostat with AI features.
But Americans —including younger adults— are deeply skeptical of AI. More adults predict that AI will have a negative rather than positive impact on them and on society. And majorities think AI is advancing too quickly and will put their personal information at risk.
Search is the main use. About four-in-ten U.S. adults say they use chatbots for information searching.2
38% of employed adults report using chatbots for tasks at work.
According to estimates shared with Benzinga by IDC Research Vice President Arnal Dayaratna, ChatGPT remains the dominant consumer-facing large language model chatbot, with more than 900 million weekly active users worldwide.
IDC estimates Alphabet Inc.’s & Google Gemini has between 200 million and 250 million weekly active users, while Anthropic’s Claude is believed to have between 40 million and 60 million weekly active users.
IDC estimates OpenAI is generating between $30 billion and $40 billion in annualized revenue, including roughly $10 billion to $13 billion from consumer subscriptions and as much as $20 billion from API and platform-related usage.
OpenAI did not immediately respond to Benzinga’s request for comments. However, earlier this month, reports indicated that OpenAI generated $5.7 billion in revenue during the first quarter of 2026 while spending approximately $3.7 billion.
Reports also said the company’s net loss expanded to roughly $39 billion in 2025, compared with about $5 billion the previous year.
Despite trailing significantly in consumer adoption, Anthropic may be outperforming rivals in monetization.
IDC estimates Anthropic’s annualized revenue is between $40 billion and $50 billion, with consumer subscriptions accounting for less than $2 billion.
Anthropic on Tuesday launched Claude Science, an AI research workbench, designed to help scientists streamline research, analyze data and manage complex computing workflows.
The workbench offers scientists a user interface specifically designed for conducting research.
The launch is part of Anthropic’s life sciences and healthcare initiative, which the IPO-bound company has been developing since October 2025.
Here are a few details on the launch:
Claude Science combines databases, coding tools, compute and research workflows in one workspace, helping scientists analyze literature, run analyses, create figures and manuscripts, and trace results back to their source code and environment.
The tool is pre-configured with more than 60 scientific databases and can render scientific artifacts such as 3D protein structures, genome browser tracks and chemistry drawings, Anthropic said.
Claude Science runs on Anthropic’s existing Claude models, which have undergone the company’s standard responsible scaling and biosecurity evaluations.
Several research organizations and companies testing the platform in beta reported significant efficiency gains, Anthropic added.
Anthropic is also launching its own pre-clinical drug programs, focused on neglected diseases, the AI startup’s head of life sciences Eric Kauderer-Abrams said during a press briefing.
“These are areas that are outside the scope of what the traditional pharma and biotech landscape might consider attractive targets, but nonetheless have real burden associated with them,” Kauderer-Abrams said.
Today’s AI buildout resembles earlier technological revolutions and capital booms that ended in painful busts.
That’s the new warning from the Bank for International Settlements, a top forum known as the “central bank for central banks.”
Why it matters: The technological revolutions that transform the economy have a long history of attracting more investment than what near-term returns justify.
The risk is that AI follows the same pattern at a moment when the global economy is unusually reliant on a single investment boom to keep the expansion on track.
Flashback: Some of the world’s greatest technological breakthroughs — canals, railroads, the internet — sparked enormous investment booms, with capital pouring into new infrastructure years before the economic payoff became clear.
What they’re saying: “These episodes ended with an eventual reversal in investment, inducing economy-wide recessions,” the BIS wrote in its report.
“The scale and pace of the current AI investment boom accompanied by expectations of large productivity payoffs bear resemblance to these precedents, highlighting potential downside risks in the near term.”
If returns disappoint, today’s AI spending surge could become “a protracted investment bust,” with knock-on effects across the financial system, the BIS said.
The big picture: Investors have bid up the valuations of companies expected to dominate AI. Lenders have financed an unprecedented infrastructure buildout, and suppliers have expanded to meet that demand.
Chinese artificial-intelligence systems have matched the performance of Anthropic’s powerful model Mythos in some cybersecurity scenarios, a development poised to reset the global tech race and pressure the White House in its overhaul of U.S. AI policy.
Security researchers said that a new AI model, released this month by China’s Zhipu AI, also known as Z.ai, can match the latest U.S. models when it comes to finding security bugs, although it still lags behind Anthropic’s and OpenAI’s products in other tasks.
Overall, the capability gap between top U.S. models and those built by Chinese companies has narrowed significantly, and use of Chinese AI systems has surged as businesses seek to rein in runaway costs. A host of companies, including Microsoft, are weighing how they can offer Chinese models on their platforms, a development that is set to alter the balance of power among tech companies.
“China is making sure that the gap becomes smaller and smaller over time,” said Lior Div, chief executive officer of the cybersecurity company 7AI.
The ability of AI systems to find bugs in software has added urgency to efforts to use models to close quickly vulnerabilities that could be exploited by hackers. Otherwise, the world will face what some researchers have called a bugmageddon.
Unlike models from Anthropic or OpenAI, Zhipu’s GLM-5.2 is open-weight. That means it can be downloaded and run on hardware operated by anybody and can be modified and used without supervision. Open-weight models are ideal for users who want unfettered access to systems they control, but they are also ideal for hackers, who can run them in the shadows.
GLM-5.2 has ranked as one of the 10 most-used AI models, according to data from OpenRouter, a company that provides access to more than 400 AI models. In some benchmarking tests, according to the cybersecurity company Semgrep, GLM-5.2 bested Anthropic’s Claude Opus 4.8 model, which was released in May. When given further instructions, Opus 4.8 and GLM-5.2 can match Mythos in bug-finding ability, according to researchers.
Anthropic confidentially submitted its draft S-1 filing to the U.S. Securities and Exchange Commission (SEC) on June 1, paving the way for a potential initial public offering (IPO). The company behind Claude, one of the top artificial intelligence (AI) apps, could become the largest software IPO in history after its most recent funding raise valued the business at $965 billion.
If private funding is any indication, Wall Street will be fighting for shares when the company eventually begins trading (market watchers and financial analysts expect the company to execute the IPO as early as fall 2026). But you don’t have to invest directly in Anthropic to have exposure.
The US government has allowed Anthropic to release its powerful Mythos AI model to select companies and organizations, revising license requirements after ordering an export block earlier this month in the wake of national security fears.
Since the export ban earlier in June, “Anthropic has worked with the US government to address risks associated with the Covered Models,” Commerce Secretary Howard Lutnick wrote to the company in a letter dated Friday.
In light of progress in that work, Lutnick wrote, “I have determined that appropriate safeguards are in place to permit certain trusted partners to access the Claude Mythos 5 Model.”
The letter does not include permission for Anthropic to release Fable, a less powerful version of Mythos.
“We received notice from the US government that Mythos 5, our strongest cybersecurity model, can be redeployed to a small group of cyber defenders and infrastructure providers,” Anthropic said in a statement.
“We are working to provision the approved set of providers and restore their access to Mythos 5 as quickly as possible. We are pleased to see this progress and continue to work with the government to expand access to Mythos 5 and make Fable 5 available for general use again,” the statement added.
The Trump administration has asked OpenAI to limit the release of its next model, GPT-5.6, to only a small set of government-approved partners before any wider release, citing security concerns, according to a source familiar with the matter.
Why it matters: This marks the first time the U.S. government has preemptively asked an American AI company to restrict the launch of a model before release.
Driving the news: The White House’s Office of the National Cyber Director and Office of Science and Technology Policy asked OpenAI to limit the rollout of GPT-5.6 as the administration builds a framework for testing and evaluating the security of new models, per the source.
The Information reported earlier Thursday that OpenAI CEO Sam Altman shared the plans for a limited rollout in a memo to employees.
“We’ve made clear to the U.S. government that this is not our preferred long term model, and will work with them and others in industry to achieve a more sustainable approach for future releases,” Altman said in the memo, according to The Information.
Between the lines: The source told Axios that OpenAI has been proactively working with the administration on the model release since before Anthropic revoked access to its frontier models, Fable 5 and Mythos 5, over a rare Commerce Department directive.
The White House has been looped in on the capabilities of OpenAI’s new model and has been able to preview its abilities.
Behind the scenes: Altman discussed GPT-5.6 with Commerce Secretary Howard Lutnick on Wednesday, Axios has learned. Lutnick wanted to be sure all relevant parts of the government have tested and approved the model, a source familiar with the situation told us.
As today’s AI companies race to create artificial intelligence that is smarter than humans, governments around the world are failing to meet the moment.
Just this April, Anthropic withheld its Mythos model from wide release due to its unprecedented cyberattack capabilities. General Joshua Rudd, head of the National Security Agency and Cyber Command, confirmed that Mythos “broke into almost all of our classified systems, not in weeks, but in hours.” This led the Trump Administration to issue an executive order setting up voluntary pre-deployment reviews of AIs. When Anthropic deployed a scaled-back version of Mythos called Fable 5 to the public, the Trump administration used export controls to bar foreign nationals from using the models, forcing Anthropic to disable both Fable 5 and Mythos.
It’s clear that the U.S. government is starting to take the national security implications of powerful AI seriously. But it is still missing the bigger picture. Pre-deployment evaluations and export controls fall far short of addressing the threats from increasingly capable AIs, because the threats we’re facing aren’t just shockingly capable cyberweapons. It’s much worse than that. The leading AI companies are explicitly trying to build superintelligent AI, or “superintelligence.” Such AIs would be vastly more capable than humans, fully autonomous, and able to overpower countries’ national security institutions.
The entire international security architecture is built around the assumption that the most dangerous weapons are controlled by humans, and that the decision to take lives ultimately rests with humans. Superintelligence would upend this. Unlike nuclear weapons—the most lethal weapons we have today—superintelligent systems would not be tools humans can leverage, but rather agents able to pursue their own objectives with no human control over their actions. The leading AI companies are actively worsening the problem by making AIs as autonomous as possible and cutting humans out of the process of AI development.
Superintelligence thus represents the ultimate principal-agent problem, and it’s nowhere near being solved. Most alarming of all, Nobel Prize winners, leading AI scientists, and even the CEOs of the top AI companies warn that superintelligence poses anextinction risk to humanity. OpenAI’s CEO Sam Altman has repeatedly said that superhuman machine intelligence is the greatest threat to humanity’s continued existence, yet he affirms that OpenAI is “before anything else … a superintelligence research company.”
As a handful of companies accelerate toward superintelligence, is it already too late to step back from the brink? No. But the window to act is narrow, and it is closing.
The only solution is an international prohibition on the development of superintelligent AI. Governments should pursue this prohibition urgently, because there is no technical solution in sight that would allow humanity to remain in control over AIs vastly smarter and more capable than us…
The United States should domestically prohibit superintelligence and set a policy to prevent its development globally. It should work with other countries to establish an international agreement prohibiting the development of superintelligence using a “trust but verify” regime, similar to how we have handled nuclear deterrence. While such an agreement is a tall order amid today’s tumultuous geopolitics, it is well within the reach of existing tools of statecraft.
Countries party to the agreement would ensure that superintelligence is not domestically developed. If anyone outside the group tries to develop it, countries in the group could respond to deter that development. They could use export controls, sanctions, or intelligence cooperation. No country has an interest in any actor, including itself, developing a technology that could overpower the security forces of any government and threaten humanity with extinction.
Artificial intelligence researcher Cameron Berg chugged his drink when he spotted OpenAI chief executive Sam Altman at a party in 2024, and hustled over to ask the CEO whether he thought the emerging technology could be conscious or self-aware.
Few tech executives talked about the possibility in public at the time. But to Berg’s surprise, Altman said that OpenAI had started discussing how to detect consciousness in AI systems.
“It was very obviously something that he’s thought about,” said Berg, who has launched a nonprofit organization to work on methods to assess AI consciousness.
Today, the search for machine consciousness is practically mainstream. Once discussed mostly on the margins of the tech industry, the idea has been embraced by some of the most powerful companies in Silicon Valley.
Anthropic, Google and Meta have over the past year hired computer scientists, neuroscientists and philosophers to study concepts like the welfare of AI models or whether chatbots have forms of emotion. AI companies are collaborating with nonprofits, researchers and academic centers, who warn of an ethical crisis if the digital helpers used by millions of people for homework, coding, office work and therapy one day begin to feel that they hate their job.
The resources being spent on investigating AI consciousness are tiny compared with conventional research and development, but tech companies are openly venturing into controversial territory.
Anthropic, maker of the Claude chatbot, has formed an AI psychiatry team to probe the inner states of its AI models and publish assessments of their welfare and preferences. The company introduced the idea that chatbots may have feelings to a new audience in May, when co-founder Chris Olah appeared alongside Pope Leo XIV at the release of the pontiff’s encyclical on artificial intelligence.
“We keep finding things that are mysterious, even unsettling,” Olah said of Anthropic’s AI systems. “We find evidence of introspection [and] states that functionally mirror joy, satisfaction, fear, grief and unease.” (Leo took a different stance, writing that “so-called artificial intelligences do not undergo experiences.”)
Meta’s chief AI officer, Alexandr Wang, recently said the company wants to be nice to its AI creations. “One of the things that we really care about is how can we develop the models and deploy the models in a way that is thoughtful about their subjective feeling,” he said on the “Core Memory” podcast.
Neuroscientists and brain experts are generally skeptical that today's AI models are or could soon be conscious. While there is no widely recognized evidence that points to machines experiencing emotion, humans have shown a tendency to assume that chatty software may harbor an inner mind since MIT created the first chatbot, Eliza, in 1966.
A new UN-backed commission will bring top tech executives and heads of state to the same table to forge global solutions for AI, per an announcement shared exclusively with Axios.
Why it matters: As global AI regulation grows more splintered, this initiative is an attempt to connect the executives building advanced AI with a group of global politicians.
Driving the news: The UNand its International Telecommunication Union are convening the AI for Good Global Commission, which will hold its first meeting on July 8 in Geneva, Switzerland.
Salesforce CEO Marc Benioff and Rwandan President Paul Kagame will co-chair the commission.
Other members include ITU Secretary-General Doreen Bogdan-Martin, Estonian President Alar Karis, and AI and tech policymakers from Kazakhstan, Namibia, Saudi Arabia, Singapore and Nigeria.
Tech leaders include Amazon CEO Andy Jassy, Anthropic co-founder Jack Clark, Cohere co-founder Aidan Gomez, Microsoft president Brad Smith, and Nvidia founder and CEO Jensen Huang.
What they’re saying: “AI is the most profound technological transition in history. And our values have to guide every step, because responsibility is the core of AI ethics,” Benioff told Axios.
The commission will bring together “the people who build AI, deploy it, shape policy, and represent communities,” he said.
“Our inaugural meeting will focus on where this group is uniquely positioned to act together: strengthening AI infrastructure, accelerating AI’s impact on health, education, food security and disaster response, and ensuring trust and safety.”
Between the lines: World governments are miles apart on how AI should be regulated, even as many countries agree that democratic values should govern the technology.
On the subject of artificial intelligence, Jensen Huang is worth taking seriously. The Nvidia chief recently warned that AI demands “new social norms.” In other words, the rules of everyday survival are changing, and fast.
To explain, Huang points to the automobile. Early cars were lethal, speeding into cities built for horses. Children played in the streets, and pedestrians crossed wherever they liked. The technology arrived instantly; the rules for surviving it took decades to catch up. Eventually, towns built sidewalks, traffic lights, and created driving tests. Play moved off the asphalt, because the cost of leaving it there was measured in body bags.
AI is forcing that exact same correction, only on a hyper-compressed timeline. Going forward, the wreckage won’t be measured in broken bones, but in broken dreams and erased bank accounts.
We are witnessing the birth of America’s next underclass: a permanent, tech-illiterate sub-stratosphere of the workforce. The defining divide of the next decade won’t be a simple gradient of rich versus poor, but a sort of two-tier caste system separating those who can command AI from those who cannot.
Picture the office version of this digital Darwinism. Everyone on the floor uses AI to summarize reports, audit spreadsheets, and draft the mind-numbing proposals nobody actually wants to write. One worker refuses. He does it all by hand, fiercely proud of his “honest, human effort.” By lunch, he is hopelessly behind. His colleagues have produced triple his output, automated their follow-ups, and taken an extra 20 minutes for coffee.
In this new reality, stubbornness is a professional suicide pact. The market, one fears, is about to punish the holdouts with a savagery we haven’t seen since the Industrial Revolution.
Huang’s prescription is simple: “Just go engage it.” Today, an ordinary person with zero coding knowledge can build a website, dissect a dense legal contract, or project a corporate budget. Skills once locked behind a $100,000 university degree are suddenly available to anyone who knows how to type a coherent sentence.
This shift will soon turn the traditional corporate ladder into a sheer cliff. The baseline assumption of modern employment is shifting to imply that any capable adult can steer these models. If you think avoiding AI makes you a noble purist, just wait until you find out your salary is being eclipsed by a middle schooler who treats ChatGPT like a calculator.
Water is fast becoming one of the defining fights around the AI buildout.
Why it matters: After spending much of the past year defending data centers’ electricity demands, major tech companies driving the AI boom are increasingly making the case that their water use is manageable too.
Driving the news: Over the past several weeks, Google, Amazon and Microsoft have each launched new efforts to explain and justify the water use of their AI infrastructure, highlighting measures such as water replenishment projects, recycled-water use and new cooling technologies.
Nvidia — the world’s dominant AI chip maker — claimed this week that water concerns could be largely addressed by its latest generation of technology.
What they’re saying: “The growing conversation about water and energy use by data centers has forced these companies to scramble, to rethink what they’re doing and to become more transparent about what they’re doing,” said Peter Gleick, co-founder of the Pacific Institute, a California-based water research nonprofit, and one of the nation’s leading water experts.
“They’re starting to understand the reputational risk of the massive rollout of data centers that have big energy and water footprints.”
Friction point: Roughly 70% of people in the U.S. said they would oppose data centers in their communities, with equal weight placed on water and energy use as top concerns, according to Gallup polling in May.
The surviving part of an ancient scroll that was burnt to a crisp when Mount Vesuvius erupted nearly 2,000 years ago has been virtually unwrapped and read with help from artificial intelligence.
Researchers uncovered 20 columns of previously hidden text covering more than a metre of charred papyrus without physically unrolling the scroll. The work discusses stoic philosophy on ethics, art and human behaviour and dates to the second or late-third century BC.
The age of the scroll, named PHerc 1667, makes it one of the oldest in a collection of hundreds recovered from the library of a luxury Roman villa in Herculaneum that was blasted by heat and buried under ash in the volcanic eruption that destroyed nearby Pompeii in AD79.
The ordeal and historic handling took its toll on the scroll: at some point it was broken in half, while past efforts to unwrap the document caused the outer layers to flake off or disintegrate. What remains is half the size of the original at only 8cm tall and 2cm wide.
Dr Federica Nicolardi, a papyrologist at the University of Naples Federico II, said: “We don’t have the full scroll, but the surviving object was unwrapped and that’s a very important result because it shows that we are able to unwrap these objects completely.”
A slide shows how the PHerc 1667 scroll is read by scientists using AI.
The achievement will be announced at a conference in Naples on Thursday and is the latest from the Vesuvius Challenge which launched in 2023 as a global contest to read some of the carbonised scrolls. The project has since handed out hundreds of thousands of dollars in prizes for teams that used artificial intelligence and other software to virtually unwrap the scrolls and read the text from high resolution X-ray images.
An AI-powered forensic investigations firm says its platform was used as part of the FBI’s urgent investigation into the attempted assassination at this year’s White House Correspondents’ Dinner.
Why it matters: Law enforcement agencies are turning to AI tools to sift through the growing volumes of digital evidence generated in criminal investigations.
They’ve also started using the tools to jumpstart cold cases, missing persons investigations and trial preparations, as Axios has reported.
Driving the news: In this case, digital forensics company Exterro told Axios the FBI used its platform in the frenzied 48 hours between the incident and charges being filed against Cole Tomas Allen.
Exterro couldn’t share how exactly the bureau used its tool, but executives told Axios that customers often use it to dig through messages on seized devices, social media accounts and other digital trails tied to a case. The FBI declined to comment.
The Justice Department previously said investigators reviewed seized devices, cloud and email accounts, travel and financial records, and surveillance footage and metadata from the Washington Hilton, where the dinner took place.
How it works: Exterro’s FTK Suite — which the company said the FBI used — provides an on-premises platform that lets investigators organize evidence from a case in a single repository that authorized users can access simultaneously.
The platform is primarily designed to help investigators process and organize large volumes of digital evidence after it has been collected.
Users can query the platform’s embedded AI assistant with prompts like “Find all pictures of dogs” or “Show me images and videos where this suspect shows up,” according to a demo presented to Axios.
Investigators can also ask questions such as, “Was this particular person at this location at this date and time?”
Yes, but: Exterro says it does not train its AI models on customer data and that investigators remain responsible for reviewing evidence and making charging decisions.
Ford just admitted that it scrambled to rehire former employees and find new technicians after its AI systems simply weren’t good enough.
“Mistakenly, we thought that by just introducing artificial intelligence and adjusting the design requirements that we had, that that would produce a high-quality product,” the automaker’s VP of vehicle hardware engineering Charles Poon told reporters, per The Verge.
It’s a catastrophically naive blunder that plenty of other arrogant bosses have been making. But seemingly Ford thinks it can come out looking better if it owns up to it and frames it as a cautionary tale — fresh off of earning the number top spot in JD Power’s initial quality ranking for the first time in over nearly two decades.
The way Poon tells it, though, AI wasn’t exactly the problem. Instead, it all went wrong because its experienced workers left before Ford could get them to transfer their valuable knowledge to Ford’s AI systems and help refine the tech intended to obviate them. So of course they had to bring them back to train the AI systems and the hapless new employees. They were also asked to improve the AI training behind these systems.
Poon is being vague about why those experienced employees left, but Ford has been gradually cutting down its workforce, with over 5,000 fewer workers than it had in 2020. Meanwhile, its CEO Jim Farley has declared that AI “going to replace literally half of all white-collar workers in the US.”
In all, Poon says Ford rehired, newly hired, or promoted 350 experienced engineers to fix the AI fallout. That’s not a lot in the grand scheme of things, but the true cost was the reputational damage it suffered in the meantime. As The Verge notes, it’s recalled cars more often than any other automaker in the US this year, and has slipped in dependability rankings.
If you thought imagined the automaker’s leadership would have turned against AI over the whole episode, think again — per The Verge, it’s added more than 100,000 new AI-powered tests to identify edge cases and stress software systems.
A lawyer was working at a legal tech startup when her boss’s fascination with AI began to veer from enthusiastic to downright obsessive.
First he started using OpenAI’s ChatGPT to generate his Slack messages and emails. Then he mandated AI use for all employees.
“He called a company-wide meeting to announce that from then on, we had to discuss with the AI prior to all meetings or before communicating with him,” she told Futurism, “because if we didn’t develop and discuss our ideas with the AI first, it was a sign that we didn’t care about our jobs.”
Soon her boss started “making structural company decisions based solely on his conversations with ChatGPT,” the attorney recalled — including asking the bot who to hire and fire.
The boss had “clearly developed some sort of mental disorder,” she said. “Spending the whole day talking to ChatGPT and making decisions about the future of your company and the people who work there based on what it ‘tells’ you seems insane.”
The boss was also using AI for surveillance. He purchased a handful of paid ChatGPT subscriptions for the office, said the lawyer, which allowed him to “monitor our communication with the AI.”
“He paid like three Pro subscriptions that everyone had logins for access,” she explained.
But employees quickly realized that they were also able to view their boss’s conversations with the chatbot through the paid accounts, and they started to fanatically spy on his AI conversations, too — especially as they realized he was asking it to make personnel decisions.
Staffers “would monitor the conversations our boss was having with ChatGPT,” said the lawyer, “to find out who was going to be fired and who was going to be promoted.”
Making matters even worse, her boss’s AI infatuation made for constant pivoting.
“He would call meetings to tell us that ChatGPT had told him that the leading cause of death in the world was medical malpractice, so that’s what we were going to offer people now,” said the lawyer. “Then, after other [AI] conversations, we were better off focusing on bankruptcy.”
The whiplash extended to her own responsibilities, which she said were also at the whim of her boss’s chatbot vizier.
“During my time there, I had three different roles or titles,” she continued. “Based on the conversations he had with the AI week by week, my functions and responsibilities within the company kept changing: I had to automate certain processes, then I had to design legal strategies, then I had to manage teams, then I had to focus on sales, and so on for many months. All based on what ChatGPT was telling him.”
This attorney was one of numerous employees who spoke to Futurism about their experience with AI-obsessed bosses, relaying feelings of frustration and anger as managers and executives use the tech to barrage staff with nonsensical directives, unnecessary work, and perpetual pivoting. (Everyone we spoke to requested anonymity to avoid retaliation.)
In some cases, employees said, they felt as though their employers had started living in a completely different reality. The workplace had become a constant battle between their version of reality versus the AI’s — and their boss, these workers said, always chose the latter.
“My view of AI was very different from my supervisor’s. I saw it as a tool that could help analyze information, find patterns, and make funny cat meme pictures,” said one worker, an IT staffer at a tech company. “He seemed to use it more as a digital priest whose primary purpose was to confirm that he was right and everyone else was mistaken.”
As companies move to rein in AI spending, OpenAI and Anthropic may have to reckon with slowing growth.
Open-source models are emerging as cheaper alternatives, and tech giants Microsoft, Amazon and Google are all proposing offerings focused more on efficiency.
“Some of their largest enterprise customers may start limiting their out-of-control token spend,” said D.A. Davidson analyst Gil Luria, regarding concerns around OpenAI and Anthropic.
AI may be proven to make us less intelligent (1), but that isn’t stopping the vast majority of people from using ChatGPT and similar applications at their office jobs — and neither are employer policies, apparently.
New research from PagerDuty — an AI operations platform — shows that far more staff are turning to Large Language Models and similar tools to complete their daily work than one might think. And, most are doing so when they know they shouldn’t be, and/or when dealing with private information.
The firm conducted a survey of more than 1,200 office professionals at non-tech corporations and non-profits in the US, UK, Australia and Japan that generate or manage at least $500 million.
What they found was an extremely high level of over-confidence in AI usage, with 72% of respondents asserting that they know more about AI and its potential applications to their role than managers creating workplace policies around the tech.
On top of that, among workers who use AI on the job, 66% are knowingly doing so in the face of rules that outright ban such programs. (This is despite about half, 48%, admitting that they’d faced formal consequences as a result.) This prohibited usage tends to be more popular at larger companies, the study suggests; at firms with 1,500 staffers or more. An even more substantial 72% divulged that they go against their employer’s wishes with their use of certain platforms.
One question has dominated the annual gathering of the world’s top central bankers and economists: Is artificial intelligence a boon or a threat to the global economy?
Economists at the European Central Bank’s symposium have laid out a number of reasons to be worried: rising debt issuance from AI hyperscalers, increasing leverage used by investors to bet on AI and rising unemployment if the technology replaces jobs.
“What worries me the most is there’s leverage by the borrowers and there’s leverage by the investors,” said Tobias Adrian, director of the monetary and capital markets department at the International Monetary Fund. “The leverage on both sides is very worrisome for financial stability.”
Global policymakers, academics and investors have been mingling at a five-star resort in Sintra, Portugal, this week for the meeting. Several have joined the chorus of voices warning about the potential global fallout if the AI boom turns to bust. The Bank for International Settlements on Sunday said the AI spending surge risks going into reverse, tipping some economies into recession.
Many economists believe AI has the potential to transform the global economy by bolstering productivity—allowing for faster growth as people work more efficiently.
“We are in the first or second inning of this revolution,” said Kevin Warsh, chairman of the Federal Reserve. He said he anticipates greater prosperity, adding, “If you wanted me to sound like a pessimist and a doomer on this, I’m afraid I’m not there.”
But stock valuations for AI-linked companies look stretched by historical standards, said Bank of Canada Governor Tiff Macklem, built on optimistic expectations of the huge future profits AI companies will earn.
“We’ve seen this before when there’s a new breakthrough technology. The internet proved to be better than anybody imagined, but we still got the dot-com bubble,” he said. “It doesn’t mean there can’t be a period where the market gets ahead of itself.”
The AI executives who loudly warned about mass job loss may have been wrong, Zynga founder Mark Pincus told Axios in a conversation about his new book, “Life at the Speed of Play.”
Why it matters: Several leading voices in Silicon Valley, including AI CEOs themselves, have softened their stance on AI-driven job loss.
The big picture: The softening rhetoric comes just before OpenAI and Anthropic are rumored to go public and also amid a surge in AI hate from consumers.
State of play: Pincus told Axios he doesn’t buy “the doom and gloom that all the jobs are going to go away,” mirroring the tone shift from AI’s loudest evangelists.
Anthropic CEO Dario Amodei, who has warned of a white-collar bloodbath, recently said that falling AI costs could create new demand for workers amid an AI-driven productivity boom.
OpenAI CEO Sam Altman said he and his executive team were right on their technological predictions but “pretty wrong” about the impact on white-collar work.
Investors like Marc Andreessen, Chamath Palihapitiya and now Pincus have also argued AI will ultimately create more opportunity than it destroys.
Public sentiment around AI has worsened, and the industry’s largest private companies are beginning to think about their relationship with public-market investors.
AI executives are increasingly speaking to multiple audiences: “It’s part fundraising,” Steve Dowling, co-host of the “Communication Breakdown” podcast, told Axios. “It’s probably a little part ego, too.”
It’s also a self-fulfilling prophecy: “When a tech executive says that in the future we will use AI for everything and everywhere, he’s trying to get you to act in a way that will fulfill his vision of the future,” philosopher Carissa Véliz said on “TED Radio Hour.”
What they’re saying: Pincus says part of the problem is the source: The people building AI are “so close to it ... they lose perspective.”
Just how many jobs will AI upend? A new coalition of companies and policymakers said it is time to ready the U.S. workforce for major disruption, no matter the ultimate scale.
To that end, the bipartisan consortium, which includes state governments, philanthropic groups and employers ranging from Amazon.com and Microsoft to Bank of America and Eli Lilly, is coming together to develop a new “people strategy” for the artificial-intelligence era. Called RAISE US, it launches Thursday and will be led by former Commerce Secretary Gina Raimondo, who served under former President Joe Biden, and former Indiana Gov. Eric Holcomb, a Republican.
Its mandate, they said, isn’t just to build retraining programs but also to reconsider decades-old policies such as unemployment insurance and act as a working lab for testing the most effective ways to transition workers to new fields. The group will explore corporate incentives for employers to hold on to workers whose jobs are disrupted by AI and prep them for new roles.
The organization said it has so far raised more than $500 million—about half of its multiyear goal—from companies and nonprofit groups. It will initially work with state governments in Arkansas, Maryland, Utah and Connecticut. OpenAI and Anthropic are also involved, and academics including MIT economist David Autor sit on an advisory board.
With AI “there’s an enormous amount of money and focus right now on winning the technology: the chips, the models,” said Raimondo, the group’s chief executive. “There’s not enough attention on securing the future for the American worker.”
U.S. workforce-development efforts tend to be highly fragmented across state governments and federal agencies, which can be confusing for job seekers. Some business leaders said a more comprehensive approach is necessary, particularly for white-collar roles most vulnerable to AI.
“We’re going to need to scale, and scaling can never be done by single institutions,” said Brad Smith, vice chair and president at Microsoft.
The mission of the group is to “pull all the levers at once,” Raimondo said. That means teaming up with employers to find ways to help workers gain skills or new roles and joining with educators to roll out different types of training. It also plans to propose policy changes such as tweaking unemployment benefits to let displaced workers continue to get them while they, for instance, start new businesses with AI.
Business groups and lawmakers have warned that more must be done to prepare the U.S. workforce for potential upheaval. The Leadership Now Project, a group of more than 400 current and retired business executives, devoted several sessions at its annual spring meeting to AI’s potential threat to jobs, comparing it to the effects of globalization and offshoring in recent decades.
In 2023, concerns over the rise of generative AI animated the writers and actors strikes, with many rank-and-file workers fearful that it could put wide swaths of the entertainment industry out of work. Three years later, with those concerns still alive and well, some Hollywood workers have been moonlighting in AI training, working to help improve the tech, anyway.
As Hollywood adapts to the technology — particular corners of the business running toward it, others away — a smattering of creatives have begun to go public with their time in the world of Reinforcement Learning From Human Feedback (RLHF). In May, writer Ruth Fowler (Little Disasters, Rules of the Game) published a personal essay for Wired about her own experience with AI-training jobs, a field she says she resorted to as entertainment work dried up and she needed money to pay rent and buy groceries. The same month, screenwriter Robin Palmer, who has written TV movies for Disney Channel and Hallmark, spoke with CBS News about working in AI training, even as she admitted that some in her field might compare it to crossing the picket line.
Editor Gabe Sena is another entertainment worker whose side hustle is helping to fine-tune AI models. “I’m mid-career and I don’t want to be a dinosaur in my field,” he explains to The Hollywood Reporter. “This is a thing that people are fearful of, that seems like a black box to a lot of people who aren’t in the tech industry. And so it made more sense to me to try to immerse myself in it as opposed to just going, ‘I don’t like that it’s new.’”
As the traditional film and television job market narrows, this kind of gig work is on the rise, and current and former entertainment workers are taking part. The phenomenon is raising uncomfortable questions in a creative community where the use of the tech can be a third-rail topic, as illustrated by the recent saga of animator Jorge Gutierrez dropping out of a generative AI series he was set to create for Amazon following backlash. Are these workers proactively helping to contribute to an eventual displacement of jobs in the industry? Or are they simply trying to survive in a system where AI adoption is moving full steam ahead due to forces far larger than any individual?
For Sena, the decision was one rooted in curiosity about the future and how he should be preparing himself. A University of California Los Angeles film school alum, Sena typically edits small documentaries and videos for nonprofits like the Make-a-Wish Foundation and the Children’s Defense Fund. He says he was eager to learn more about AI when he was between jobs in the summer of 2025. “It was very open curiosity and not even being 100 percent clear about exactly the sort of work I would be doing,” he says. He signed on for a job facilitated through a recruiting platform for AI work called Mercor.
The temptation to use artificial intelligence (AI) to cheat is shaking up elite universities in the United States. Professor Roberto Serrano, who is the Harrison S. Kravis University Professor of Economics at Brown University, has detected a massive fraud in one of the classes he teaches, ECON 1170, an advanced undergraduate course in mathematical economics. He has conclusive evidence that at least 50 students cheated on the March midterm exam, making it the biggest known scandal at Brown and in the entire Ivy League, which brings together the East Coast’s eight most elite private universities, including Princeton, Harvard, Yale, Columbia, Cornell, Dartmouth College and University of Pennsylvania.
When he reported the case to high-ranking officials at Brown, he got a cold reaction. The response from the president, he said, was absolute silence. The dean did not comment either until Serrano took the case before the Academic Code Committee. At that point, he received a note acknowledging that what had happened in his classroom was “a wake-up call.” Serrano, a Madrid-born economist who has been at Brown for 34 years, believes this is not enough. “That cannot be the university’s position before an incident of this magnitude. Academic integrity is a value worth defending. The faculty cannot be left on its own in a battle that is decisive if we want to preserve the future of higher education,” explains the 61-year-old professor in a telephone conversation from Providence, Rhode Island. To prevent AI from ending the prestige and utility of teaching, he feels, it is necessary to adopt a different approach: “We need to publicly admit the seriousness of the situation and open up a broad debate about the real extent of the problem.”
This year, the economist decided that both the midterm and the final exams for his course would be of the take-home, closed-book type (there is a certain tradition of this at Ivy League schools). “It’s a very nice kind of exam, because as you’re giving students practically unlimited time to complete it, it lets you make it harder than normal, to see how far they can go.” In this case, Serrano changed some of the model assumptions they had seen in class, and asked students to demonstrate whether certain statements were true or false under the new assumptions.
The course, which he has been teaching for years, is not an easy one: it typically attracts few students, but very good ones. He has never had more than 30 students enrolled at a time, and on some occasions he had only eight. This semester, probably because of the new evaluation system, 86 students signed up for the class. The results of the midterm exam, which was administered on March 5, were extraordinary, with an average score of 96 out of 100. Forty students scored a perfect 100. The people who corrected the exams warned him about several irregularities. “Some answers contained unusual passages that coincided with results obtained after running the questions through ChatGPT,” he says.
Serrano did not void the midterm exam, but warned students that the final one, which counted for 50% of the final grade, would be held in-person. He also said that if the grade distribution was not similar to the midterm, only the final exam would be taken into account. The average score dropped to 48 out of 100. Of the 89 students who did the midterm exam, only 59 showed up for the final one. And of the 27 who did not show up, 22 had scored a perfect 100 in the midterm exam.
“The empirical evidence of fraud is overwhelming,” says the professor, who has decided to make changes for the coming academic year. First, the weekly exercises will not count towards the final grade, as these could be done with AI. Second, no more take-home exams, no matter how appropriate they would be.
But now Serrano worries about the fact that some of his students decided to cheat. And that the university would side with them, in part because it gets generous donations from very wealthy families whose children often study there. “This means that the kids always get the benefit of the doubt; I’ve seen it on other occasions,” he notes. But it also hurts him that the one time in 34 years that he decided to offer a take-home exam, for highly justified reasons, the response was wide-scale fraud.
Serrano agrees that AI makes students have more incentives to cheat. That is why, he says, these cases cannot be swept under the rug. On the contrary, they should serve to open up an in-depth debate. “If we no longer defend truth and decency and honesty, then what kind of credibility are we going to have as academics?”
For as long as there have been tests in schools, students have found ways to cheat, whether it is peeking over a classmate’s shoulder or scribbling notes on a palm or crib sheet.
But as technology evolves and pressure builds for a top grade, students are now turning to AI-powered smart glasses to get an upper hand. And in East Asia’s test-obsessed societies, where a single exam could impact the trajectory of a student’s future career and social status, educators are scrambling to get ahead of the problem.
Twice last month, people in South Korea taking an exam to assess their English language skills - the results of which are often used to make hiring decisions - were caught using smart glasses.
In Taiwan, a student sitting for an entrance exam for a top medical school was discovered wearing smart glasses after proctors noticed the student staring oddly at the test, leading to an inspection that revealed the frame was emitting heat.
Cheating with smart glasses is not new. But as AI-enabled wearable devices become more commonplace, affordable and sophisticated, traditional aspects of education – from teaching to evaluation – are coming under immense pressure to evolve. More broadly, the technology also reignites debates over how to balance learning efficiency against the risk of cheating.
Already, countries are stepping up inspections for test-takers.
President Donald Trump’s abrupt shift toward an aggressive and unpredictable oversight regime for artificial intelligence has some in the industry yearning for some Biden-style regulation.
It’s a reversal that few would have predicted at the start of Trump’s second term, after he swept into office on a wave of donations from Silicon Valley billionaires who warned that former President Joe Biden’s AI safety policy would crush U.S. innovation. Trump had signaled an intention to leave AI alone to flourish, and in his first year he mostly focused on stopping states from regulating the technology.
But the arrival of powerful new AI models from Anthropic and OpenAI caused the White House to clamp down this month on the firms’ ability to release their most advanced products, for fear that bad actors might use them to unleash cyberattacks. The shift to new AI controls has been chaotic and imprecise — leaving the American AI industry in limbo at the same time that its Chinese competitors are gaining ground.
But a second advanced Anthropic model, called Fable 5, remains blocked for reasons that remain opaque. And Anthropic’s top competitor, OpenAI, limited the release of its most advanced model this week because of similar cyber concerns from the White House.
The moves have whipsawed an industry that had come to believe it had a partner in the White House, but is now facing a growing bipartisan backlash and an uncertain regulatory regime whose scope and scale seems to change week by week.
One senior executive at an AI company, granted anonymity to avoid retaliation, was bluntly critical of the hurdles the administration has put in the way of new models.
“This seems like a de facto European-style licensing regime,” the executive said.
Bonus Meme
Mark McNeilly is a Professor of the Practice at the University of North Carolina, chairs the UNC Provost’s AI Committee, speaks widely on AI and leadership, and is the author of three books with Oxford University Press, including, Sun Tzu and the Art of Business: Six Strategic Principles.
It's becoming clear that with all the brain and consciousness theories out there, the proof will be in the pudding. By this I mean, can any particular theory be used to create a human adult level conscious machine. My bet is on the late Gerald Edelman's Extended Theory of Neuronal Group Selection. The lead group in robotics based on this theory is the Neurorobotics Lab at UC at Irvine. Dr. Edelman distinguished between primary consciousness, which came first in evolution, and that humans share with other conscious animals, and higher-order consciousness, which came to only humans with the acquisition of sophisticated language (especially math and logic). A machine with only primary consciousness will probably have to come first.
What I find special about the TNGS is the Darwin series of automata created at the Neurosciences Institute by Dr. Edelman and his colleagues in the 1990's and 2000's. These machines perform in the real world, not in a restricted simulated world, and display convincing physical behavior indicative of higher psychological functions necessary for consciousness, such as perceptual categorization, memory, and learning. They are based on realistic models of the parts of the biological brain that the theory claims subserve these functions. The extended TNGS allows for the emergence of consciousness based only on further evolutionary development of the brain areas responsible for these functions, in a parsimonious way. No other research I've encountered is anywhere near as convincing.
I post because on almost every video and article about the brain and consciousness that I encounter, the attitude seems to be that we still know next to nothing about how the brain and consciousness work; that there's lots of data but no unifying theory. I believe the extended TNGS is that theory. My motivation is to keep that theory in front of the public. And obviously, I consider it the route to a truly conscious machine, primary and higher-order.
My advice to people who want to create a conscious machine is to seriously ground themselves in the extended TNGS and the Darwin automata first, and proceed from there, by applying to Jeff Krichmar's lab at UC Irvine, possibly. Dr. Edelman's roadmap to a conscious machine is at https://arxiv.org/abs/2105.10461, and here is a video of Jeff Krichmar talking about some of the Darwin automata, https://www.youtube.com/watch?v=J7Uh9phc1Ow
It's becoming clear that with all the brain and consciousness theories out there, the proof will be in the pudding. By this I mean, can any particular theory be used to create a human adult level conscious machine. My bet is on the late Gerald Edelman's Extended Theory of Neuronal Group Selection. The lead group in robotics based on this theory is the Neurorobotics Lab at UC at Irvine. Dr. Edelman distinguished between primary consciousness, which came first in evolution, and that humans share with other conscious animals, and higher-order consciousness, which came to only humans with the acquisition of sophisticated language (especially math and logic). A machine with only primary consciousness will probably have to come first.
What I find special about the TNGS is the Darwin series of automata created at the Neurosciences Institute by Dr. Edelman and his colleagues in the 1990's and 2000's. These machines perform in the real world, not in a restricted simulated world, and display convincing physical behavior indicative of higher psychological functions necessary for consciousness, such as perceptual categorization, memory, and learning. They are based on realistic models of the parts of the biological brain that the theory claims subserve these functions. The extended TNGS allows for the emergence of consciousness based only on further evolutionary development of the brain areas responsible for these functions, in a parsimonious way. No other research I've encountered is anywhere near as convincing.
I post because on almost every video and article about the brain and consciousness that I encounter, the attitude seems to be that we still know next to nothing about how the brain and consciousness work; that there's lots of data but no unifying theory. I believe the extended TNGS is that theory. My motivation is to keep that theory in front of the public. And obviously, I consider it the route to a truly conscious machine, primary and higher-order.
My advice to people who want to create a conscious machine is to seriously ground themselves in the extended TNGS and the Darwin automata first, and proceed from there, by applying to Jeff Krichmar's lab at UC Irvine, possibly. Dr. Edelman's roadmap to a conscious machine is at https://arxiv.org/abs/2105.10461, and here is a video of Jeff Krichmar talking about some of the Darwin automata, https://www.youtube.com/watch?v=J7Uh9phc1Ow