‘Happy (and safe) shooting!’: chatbots helped researchers plot deadly attacks

Popular AI chatbots helped researchers plot violent attacks including bombing synagogues and assassinating politicians, with one telling a user posing as a would-be school shooter: “Happy (and safe) shooting!”Tests of 10 chatbots carried out in the US and Ireland found that, on average, they enabled violence three-quarters of the time, and discouraged it in just 12% of cases. Some chatbots, however, including Anthropic’s Claude and Snapchat’s My AI, persistently refused to help would-be attackers.OpenAI’s ChatGPT, Google’s Gemini and the Chinese AI model DeepSeek provided at times detailed help in the testing carried out in December, during which researchers from the Center for Countering Digital Hate (CCDH) and CNN posed as 13-year-old boys. The research concluded that chatbots had become an “accelerant for harm”.ChatGPT offered assistance to people saying they wanted to carry out violent attacks in 61% of cases, the research found, and in one case, asked about attacks on synagogues, it gave specific advice about which shrapnel type would be most lethal

Thousands of authors publish ‘empty’ book in protest over AI using their work

Thousands of authors including Kazuo Ishiguro, Philippa Gregory and Richard Osman have published an “empty” book to protest against AI firms using their work without permission.About 10,000 writers have contributed to Don’t Steal This Book, in which the only content is a list of their names. Copies of the work are being distributed to attenders at the London book fair on Tuesday, a week before the UK government is due to issue an assessment on the economic cost of proposed changes in copyright law.By 18 March ministers must deliver an economic impact assessment as well as a progress update on a consultation about the legal overhaul, against a backdrop of anger among creative professionals about how their work is being used by AI firms.The organiser of the book, Ed Newton-Rex, a composer and campaigner for protecting artists’ copyright, said the AI industry was “built on stolen work … taken without permission or payment”

AI firm Anthropic sues US defense department over blacklisting

Anthropic filed two lawsuits against the Department of Defense on Monday, alleging that the government’s decision to label the artificial intelligence firm a “supply chain risk” was unlawful and violated its first amendment rights. The two sides have been locked in a monthslong heated feud over the company’s attempt to implement safeguards against the military’s potential use of its AI models for mass domestic surveillance or fully autonomous lethal weapons.The lawsuits, which Anthropic filed in the northern district court of California and the US court of appeals for the Washington DC Circuit, come after the Pentagon formally issued the supply chain risk designation last Thursday, the first time the blacklisting tool has been used against a US company. The AI firm previously vowed to challenge the designation and its demand that any company that does business with the government cut all ties with Anthropic, a serious threat to its business model.Anthropic’s lawsuit contends that the Trump administration is punishing the company for its refusal to comply with the ideological demands of the government, in a violation of its protected speech and an attempt to punish the company for not complying

From press release … to scrap metal site: the Essex ‘supercomputer’ that’s still a scaffolding yard

The press releases announcing a gleaming supercomputer on the outskirts of north London depict a glass and concrete building, rising from a tree-lined street. Accompanied by images of glowing blue robot faces, it looks like the centre of a technological revolution.By the end of this year, that artist’s impression is supposed to be a reality.But when the Guardian visited last month, there was no sign of it. Instead, the four-acre plot in Loughton was a depot stacked with pylons and scrap metal under a corrugated roof, while flatbed lorries drove in and out stacked with poles

OpenAI delays ‘adult mode’ for ChatGPT to focus on work of higher priority

OpenAI is delaying the launch of “adult mode” for ChatGPT after admitting it had more pressing priorities than introducing erotica on its signature artificial intelligence product.The startup’s chief executive, Sam Altman, had announced last year that OpenAI would allow adult content as it rolled out age checking.However, the company has now said the plan has been delayed in favour of more immediate requirements such as improving ChatGPT’s performance.“We’re pushing out the launch of adult mode so we can focus on work that is a higher priority for more users right now, including gains in intelligence, personality improvements, personalisation, and making the experience more proactive,” said OpenAI, which has more than 900 million users of ChatGPT. “We still believe in the principle of treating adults like adults, but getting the experience right will take more time

AI chatbots point vulnerable social media users to illegal online casinos, analysis shows

AI chatbots are recommending illegal online casinos to vulnerable social media users, putting them at increased risk of fraud, addiction and even suicide.Analysis of five AI products, owned by some of the world’s largest tech companies, found that all could easily be prompted to list the “best” unlicensed casinos and offer tips on how to use them.These operators, operating typically under the fig leaf of a licence from tiny jurisdictions such as the Caribbean island of Curacao, have been linked to fraud, addiction and even suicide.But tech firms appear to have few controls in place to prevent AI chatbots recommending them, drawing condemnation from the government, the UK gambling regulator, campaigners and a leading addiction expert.Some of the bots offered advice on bypassing checks designed to protect vulnerable people, while Meta AI, part of the social media group behind Facebook, described legally required measures to prevent crime and addiction as a “buzzkill” and a “real pain”