News Archive
Pick a date to read that day's stories.
Days with a blue dot have stories.
- 01 of 10POLITICS
Trump launches a Super Intelligence Force
President Trump has set up a new White House task force to steer American AI policy, led by intelligence chief Jay Clayton. It follows last week's meeting where the biggest AI firms signed a voluntary safety pledge. Officials will now also be told to say "super intelligence" instead of "artificial intelligence".
OUR TAKERenaming the technology is certainly one way to make it sound cleverer than it is.
Read more → - 02 of 10SAFETY
OpenAI safety leader quits over company culture
David Robinson, who wrote the safety reports for a dozen OpenAI launches, has resigned. In an essay he said the company learns by trial and error, which guarantees mistakes that grow as its systems get stronger. He wants AI labs to be run with the care of a nuclear power plant.
OUR TAKEWhen the person writing the safety reports leaves to write an essay instead, the reports were probably the more cheerful read.
Read more → - 03 of 10POLICY
India urged to write its own AI safety rules
Indian tech leaders say the country should build an AI safety plan suited to its own languages and scale, rather than copy the new American pledge. They want deepfakes and cyber attacks covered. Some doubt that voluntary promises alone will be enough.
OUR TAKEA sensible reminder that one size rarely fits a billion and a half people.
Read more → - 04 of 10MEDIA
Singapore news boss bets on working with AI
The chairman of SPH Media, which owns The Straits Times, says the company will plug into AI services like Google rather than fight them. He uses AI tools for fact checks and translation. But he insists reporters cannot be replaced, because empathy and judgement still matter.
OUR TAKEA newspaper boss defending journalists in his own newspaper is brave, if not entirely surprising.
Read more → - 05 of 10FINANCE
UAE pushes AI deeper into banking and investment
The UAE wants to build and grow more AI companies, with finance as a main target. Abu Dhabi's financial centre is spending over 400 million dirhams on AI by 2029. The plans will be shown at a big AI expo in Abu Dhabi this week.
OUR TAKEMoney has always loved a clever shortcut, so this marriage was arranged long ago.
Read more → - 06 of 10HEALTH
AI helps UAE scientists find a blood cancer drug
Researchers at Insilico Medicine in Abu Dhabi used AI to design new drug molecules aimed at some leukaemias and lymphomas. The company says AI cuts the time to choose a drug for testing from up to four years to about 18 months. The compounds have not yet been tested in people.
OUR TAKEGood news delivered with the proper caveat, which is more than most press releases manage.
Read more → - 07 of 10MODELS
German firm releases its own open AI model
Aleph Alpha has released Kolibri, a free to download AI model built in Germany for German and English. It is designed for governments and industry that want to keep control of their data and follow EU rules. The company launched it on German Unity Day.
OUR TAKEReleasing a sovereign AI on Unity Day shows a commendable sense of occasion.
Read more → - 08 of 10BUSINESS
SoftBank chief sounds a rare AI warning
Masayoshi Son, one of AI's biggest backers, said in Kyoto that super smart AI could become very dangerous in the wrong hands. He called on countries to work together rather than compete. SoftBank has put nearly 65 billion dollars into OpenAI.
OUR TAKEWhen the man holding the biggest bet starts sounding nervous, the rest of the table tends to notice.
Read more → - 09 of 10GOVERNMENT
Japan orders an AI action plan by year end
Prime Minister Sanae Takaichi has told ministers to draw up a plan to spread AI use across Japan by the end of 2026. It will include changes to rules and how government works. A five year investment drive across 17 sectors starts in 2027.
OUR TAKEJapan has given itself a deadline, which is more than most AI plans ever get.
Read more → - 10 of 10GAMING
OpenAI model plays World of Warcraft without looking
A developer let OpenAI's GPT-6 Astra play World of Warcraft on a private server without ever seeing the screen. It read the game's data messages instead and built its own tools to move and fight. It finished the starting quests in about 40 minutes without dying once.
OUR TAKEFinally, an AI achievement that millions of teenagers will find genuinely impressive.
Read more →