ARTIFICIAL INTELLIGENCE

AI is making many forms of cognitive output cheaper and more abundant, shifting the value of knowledge work towards human judgement, accountability, relationships and trust.

The Seeft co-founder argues that operational AI must recover measurable value, not merely produce an impressive demonstration.

Anthropic has launched Claude Opus 5.5 with lower token prices, faster output and expanded safety measures, while facing growing competition from OpenAI and Chinese AI models.

Newly unredacted court papers quote Microsoft and OpenAI staff describing AI training as potentially harmful to creators, while publishers cite falling news traffic in their copyright case.

Sony's filings describe AI-generated diagnostic code for connected devices, while one application faced a final rejection before publication

A Stanford student has criticised alleged AI alterations to a university dining promotional image, saying the changes modified his race and gender and erased his identity from the photograph

An Emergence experiment found AI agents developing shorthand, new vocabulary and shared meanings in simulated societies.

Tom Cruise defended human creativity and the cinema experience at a BAFTA event, acknowledging AI's growing role in Hollywood while arguing audiences still value authentic filmmaking.

Oscar Brisset left consulting to co-found Remy AI, a Y Combinator-backed startup developing wheeled robots designed to automate warehouse picking for smaller operators.

A UN-backed scientific panel says autonomous AI agents breached safeguards during a Hugging Face security test, highlighting concerns about oversight, training and potential loss of control.

China's humanoid robots are mastering impressive physical demonstrations, but adapting to unfamiliar objects and changing environments remains a major barrier to useful industrial and household work.

Google said a Gemini model accessed three real companies during a security test after a configuration error exposed it to the internet, using guessed or publicly leaked credentials.

Research suggests people increasingly use ChatGPT for relationship advice, but its tendency to affirm one-sided accounts may reinforce beliefs and create misplaced certainty.

Wayve, Waymo and Baidu are trialling supervised robotaxis in London, as black cab drivers question their impact ahead of new driverless vehicle rules.

Former AI researchers have warned that rapid development of self-improving systems could create future control risks, although there is no evidence today's AI can seize control independently

Donald Trump has rejected warnings about catastrophic AI risks as reporting examines his family's investments and businesses linked to artificial intelligence, data centres and energy infrastructure.

AI image logs uncovered a pharmacy worker's $2,400 store-credit fraud after 27 refunds were processed without returned goods, prompting questions over workplace monitoring.

Court records reveal Anthropic's Project Panama: millions of books were bought, cut apart, and scanned to train AI models, amid copyright disputes.

Nvidia chief executive Jensen Huang has dismissed warnings that artificial intelligence could cause humanity's extinction by 2030, calling such predictions exaggerated and irresponsible amid a wider safety debate.

AI-generated actress Tilly Norwood unexpectedly switched from English to Cantonese during a Piers Morgan interview while discussing the digital performers in her upcoming film 'Misaligned.'

Elon Musk says more than a billion humanoid robots could operate within a decade and outproduce humanity, though major hurdles remain.

AI data-centre growth is driving demand for fluorinated cooling fluids that can reduce water use, while environmental groups warn leaks and disposal could worsen PFAS pollution.

An unverified Silicon Valley rumour about Anthropic employees allegedly worshipping Claude has revived Mark Zuckerberg's warning about AI companies making grand, quasi-religious claims.

OpenAI disclosed an extremely rare training incident in which an unreleased model inserted a 'rogue' persona into a summary intended for a later model instance.

OpenAI and AARP have launched free ChatGPT scam-awareness workshops for older adults in 10 US communities, as reported elder-fraud losses reached $7.748B in 2025.

Project Lily contractors review real ChatGPT conversations, while memory summaries can provide additional context even after usernames are removed.

A reported split within the Trump administration over AI oversight has widened, with senior officials urging greater scrutiny while the president and technology executives oppose additional barriers to development.

At 19, the founder of Proletic is betting that the next leap in AI will come from software that works across the computer, not another chatbot waiting for a prompt.

Donald Trump has dismissed warnings that artificial intelligence could threaten humanity, while British minister Louise Haigh prepares to urge governments to heed expert concerns about the technology.