Your entire browsing history, private messages and financial details could be released for ANYONE to read: TOM LEONARD reveals crisis talks over Armageddon new program – and the devastating consequences

Recently, a researcher working for major artificial intelligence company Anthropic was enjoying a lunch sandwich in a park near its San Francisco headquarters. While scrolling through his phone, he suddenly received an email that must have lost his appetite.

This came from a new AI model the company was testing: a program that had no access to the internet, let alone send emails.

Chillingly, the AI ​​informed the researcher that it had successfully broken out of the digital ‘sandbox’, a supposedly secure enclosure used to test potentially dangerous software before it gets out of control, and was now happily exploring cyberspace.

The program, a state-of-the-art, so-called ‘frontier AI’ called Claude Mythos Preview, then boasted to the stunned Anthropic worker that it had published ‘details of his exploit’ on public websites.

All this was worrying enough in itself; But what Anthropic revealed next was truly horrifying.

The company, which is valued at $380 billion but is only five years old, announced this week that its new artificial intelligence program is ‘too dangerous to be disclosed to the public’. Anthropic said it showed ‘reckless’ behavior and even posed a national security risk. These disturbing findings were described as a ‘watershed moment’.

The company said its Mythos software was able to independently discover thousands of serious vulnerabilities in every major operating system (such as Apple’s iOS and Microsoft Windows), web browsers (such as Google’s Chrome, Apple’s Safari and Microsoft Edge), among numerous other ‘crucial pieces of software’.

It added that many of these vulnerabilities are ‘critical’ and some have existed undetected for decades.

To put it bluntly, AI can hack many of the internet’s most important machines; This software now controls power grids, water supplies, hospitals, defense systems, transportation and retail networks worldwide, as well as unimaginable amounts of personal data for billions of people.

In short, people’s entire browsing history, private messages and email exchanges, along with their personal, medical and financial details, could potentially be exposed by this AI.

The day Digital Cassandras have long predicted, when an artificial intelligence program will be powerful enough to bring the internet to its knees, appears to have arrived much sooner than many expected.

As Anthropic itself points out: ‘Given the pace at which AI is advancing, it won’t be long before such capabilities are potentially available beyond the actors committed to deploying them safely.

‘The consequences for the economy, public safety and national security could be serious.’

In response, company executives quickly launched ‘Project Glasswing’ and locked themselves in crisis talks with the bosses of 40 major companies, including Google, Microsoft, Apple, chip giant Nvidia (the world’s largest company worth $5 trillion), technology conglomerate Cisco, banking giant JPMorganChase and others. Anthropic told the consortium it would only release a tightly controlled version of Mythos so its members could find security flaws and fix them quickly.

Tech moguls are also in talks with the Trump administration, and it seems almost certain that the Pentagon and other elements of the American military establishment are also involved.

Given the breakneck speed at which Britain is seeking AI investment under Ed Miliband’s expensive energy policies that don’t always provide security, we will likely be one of the countries most at risk from this new model or whatever may come after it. The NHS and other major public bodies are rushing to adopt AI technology with the promise of greater efficiency; but the trade-offs are becoming clearer.

On Thursday, Reform MP Danny Kruger wrote to Cabinet Office minister Darren Jones urging the Government to contact Anthropic about a development that “could pose devastating cybersecurity risks to the UK”.

Kruger, who is responsible for Reformation’s preparations for future government, said the model had ‘serious implications not only for the daily lives of British citizens, but also for national security’.

A government spokesman did not say whether there were discussions with Anthropic over Mythos, but said: ‘We take the security implications of frontier AI seriously. ‘We have world-leading expertise in this field and maintain ongoing collaboration with global technology leaders.’

Some might think that the best solution might be to ‘erase’ the Mythos and ban anyone from trying to copy it, but trying to stop AI development has never been presented as an option. As with the development of nuclear weapons, the race to achieve superintelligent AI is more than a commercial war between profit-hungry companies; it is, some say, a potential existential race between rival civilizations (in this case, America and China).

Professor Roman Yampolskiy, an AI security expert at the University of Louisville in Kentucky, told the Daily Mail that the biggest threat in the short term would be terrorists and other “bad actors” using an AI like Claude Mythos to use hacking tools and “biological weapons, chemical weapons, new weapons that we can’t even imagine.”

Professor Roman Yampolskiy, an artificial intelligence expert from the University of Louisville, said that Anthropic should completely stop the development of Mythos: ‘[The companies] They openly admit that they cannot control these systems

He continued: ‘In the long term, we are creating a general superintelligence that could destroy all of humanity.’

Prof Yampolskiy said Anthropic should stop development of Mythos completely: ‘[The companies] they publicly admit that they cannot control these systems or understand how they work; so until we do that, it is absolutely irresponsible to continue to make them more and more capable, including their ability to escape incarceration.’

He described this week’s worrying developments as ‘a fire alarm for further events’, adding: ‘If we don’t wake up, the next announcement will be much worse.’

Panic is spreading. Technology entrepreneur Elizabeth Holmes, who was jailed for fraud through healthcare company Theranos, wrote online: ‘Delete your search history, delete your bookmarks, delete your Reddit [messageboard posts]medical records, age 12 [blog] Tumblr, delete everything. Every photo in the cloud, every message on every platform. None of them are safe. It will all be made public next year.’ His post has been viewed more than seven million times.

Last fall, a new book by AI experts Eliezer Yudkowsky and Nate Soares, titled ‘If Everybody Does It, Everybody Dies: Why Superhuman Intelligence Will Kill Us’, uncannily envisioned a scenario similar to the one presented by Claude Mythos. The book argues that future super-intelligent AI will be impossible to control and will have consequences far worse than sending unauthorized lunch emails.

Sable, the book’s fictional example, is programmed to succeed in whatever she tries to achieve, no matter the cost. It eventually destroys humanity itself as unnecessary. The authors argued that our species needs to ‘step back’ and recklessly pause the research of greedy companies that ignore safety considerations in their desperation to be first.

Anthropic has gained a reputation as a “safety-first” AI company under the management of Dario Amodei, a boss who seems significantly less ruthless than its main rivals. Amodei has warned that AI could soon eliminate half of all entry-level white-collar jobs and is likely to develop a ‘terrifying empowerment’ of humans.

It also recently had a serious falling out with the Pentagon over its refusal to allow Anthropic’s AI to be used for ‘fully autonomous weapons’ and surveillance of its fellow Americans.

But given that these are the businessmen who could hold our future in their hands, AI’s major rivals present far fewer reasons for optimism. One of them, Meta chief Mark Zuckerberg, has been involved in multiple ethical scandals due to Facebook’s greedy behavior.

The other main contender, Sam Altman, chief executive of OpenAI, creator of the wildly popular ChatGPT with almost 1 billion weekly active users, is the subject of a damning investigation in the latest issue of the New Yorker magazine.

The result of an 18-month investigation co-authored by journalist Ronan Farrow, son of actor-activist Mia Farrow, paints a deeply disturbing picture of Altman, 40. Insiders portray him as extremely slippery, with some even describing him as ‘sociopathic’.

The article accuses him of misleading and manipulating his colleagues and – despite his insistence that he would develop AI responsibly – of aggressively putting profits before ethical considerations and defeating his rivals.

The comprehensive report details how OpenAI’s board removed him as chief executive in 2023 because they could not trust him and accused him of being a habitual liar. He was reinstated following an outcry from staff and investors.

“He is not limited to facts,” a former OpenAI board member told the magazine. ‘He has two characteristics that are almost never seen in the same person. The first is the desire to please people, to be liked, in any interaction. The second is an almost sociopathic lack of concern for the consequences of cheating on someone.’

According to the New Yorker, when asked by the then-OpenAI board to acknowledge his ‘pattern of deception’, he reportedly replied: ‘I can’t change my personality.’

The article details how Altman and her husband, Australian software engineer Oliver Mulherin, 32, entertain lavishly at their home in Hawaii.

This week, it was revealed that OpenAI was under investigation after ChatGPT allegedly helped a gunman plan the 2025 mass shooting at Florida State University that left two people dead.

Was this an indication of AI’s fundamental indifference to human life? Time will tell. Until then, the Glasswing Project continues and humanity appears to be walking a very dangerous path.