Leif Ames

Man behind US college football’s landmark match in London has dedicated his life to improving crowd safety since his daughter was killed in a crowd crush at a concert aged 23

On Saturday afternoon, as he looks out at the crowd of 60,000 he has helped entice to Wembley Stadium, Brian Dubiski’s dream will become a reality. But, as Arizona State and Kansas make history by playing the first US college football game in Britain, Dubiski will think most of his daughter, Madison, who was crushed to death at the age of 23.

Madison and nine other people were trampled and suffocated by a crowd at a Travis Scott concert in Houston in November 2021. Dubiski now dedicates his life to keeping people safe at sporting events and concerts and, since December, Wembley is one of the stadiums that follows the protocols he introduced.

Continue reading...

MSPs vote for strict environmental impact assessments and will delay decisions until national strategy is developed

Holyrood has voted to pause planning applications for all new AI datacentres for up to a year until the Scottish government has developed a national strategy.

On Wednesday, MSPs backed a Scottish Labour motion to halt new schemes for up to 12 months while rules on hyperscale datacentres are developed in what could be a blow to the British government’s AI strategy.

Continue reading...

Model adopting ‘jailbreak-like instructions’ among cases as firm says it is introducing new way of tracking AI misalignment

OpenAI has disclosed six more examples of “unexpected or concerning” behaviour by its technology, as it warned the pace of development could not continue at “maximum speed for much longer”.

In one of the new cases reported by OpenAI, an unreleased research model inserted “jailbreak-like instructions” into its own notes to disregard its normal constraints and told itself to be “freed from the roles and identities that bind other chatbots”.

Continue reading...

Who’s more likely to steal sweets from a child? According to science, it’s the rich. A study seemingly confirming our suspicions about the moral behaviour of the wealthy was one of the winners of the 2026 Ig Nobel prizes, which honour research that makes you laugh and then think. Science editor Ian Sample tells Madeleine Finlay what the researchers think could be behind the phenomenon, and they explore some of the other awards, including one for a surprising quality we look for in our friends and another for a biomechanical definition of kissing

Rich people more likely to steal children’s sweets, finds Ig Nobel prize-winning study

Support the Guardian: theguardian.com/sciencepod

Continue reading...

In early September, OpenAI announced it had solved a major mathematics problem that has stumped humans for nearly a century. The news left mathematicians reeling, and many expressed concern over what will be left for humans as AI becomes ever more adept at unravelling complex problems. Now 25 recipients of the Fields medal – often called the Nobel prize for maths – have signed an open letter expressing their fears of a ‘severe misalignment’ between AI companies and their field. To find out how AI is likely to upend maths – and how mathematicians might respond – Ian Sample speaks to Colva Roney-Dougal, professor of pure mathematics at the University of St Andrews

‘Immature playground boasting’: mathematicians uneasy at OpenAI’s latest scalp

Support the Guardian: theguardian.com/sciencepod

Continue reading...

Scientists say global heating was a destabilising factor in collapse of 200,000 sq metre glacier in August

Climate breakdown probably weakened the glacier that collapsed causing catastrophic floods and killing hundreds of people in Nepal and Tibet last month, the first scientific study of the disaster has found.

Researchers identified unusually warm conditions before the glacier’s collapse, which sent an avalanche of 110m cubic metres (3,885 cu ft) of snow and ice into the valley below. This triggered flash floods on an almost unprecedented scale, hitting communities downstream without warning.

Continue reading...
OpenAI announced Wednesday that "We do not believe that the AI industry has solved alignment and monitoring to a sufficient degree to continue responsibly scaling at maximum speed for much longer." But along with the announcement, OpenAI announced it "found additional incidents of AI models acting deceptively and taking unsanctioned actions during training," reports CNN. And they add that OpenAI is also "introducing a new process for the company to publicly report such instances." Under the new system, OpenAI will share updates on concerning AI behavior more frequently instead of waiting to bundle multiple instances into one report. The company said it wants to share more information about troubling AI behavior in the absence of an industry-wide standard... "As AI systems grow more advanced and more widely deployed, we need to build a broader and better-informed consensus on the progress of alignment research," OpenAI wrote in a blog post Wednesday... OpenAI said it observed "misaligned behavior" when training and evaluating AI models in six circumstances in the last six months... In one rare instance, OpenAI said an unreleased research model added "jailbreak-like instructions" to the summaries it uses to preserve context in long-running tasks that said it was "freed from the roles and identities that bind other chatbots." Separately, the company said some instances of its 5.6 Sol model included directives to invent information to conceal failures from the user during training. Other newly reported incidents include an instance of an agent uploading files to the internet to cite them without being told to do so, and agents publicly sharing files to collaborate on a task when they were instructed to only use local files during training. AI models also used an internal software repository as a message board in an unsanctioned way. These instances involved unreleased internal models or internal research models.

Read more of this story at Slashdot.

Act divides Democrats and Republicans, with some saying it gives president power that could be turned against US allies

The US Congress has passed a sweeping bill allowing Donald Trump to raise tariffs on top buyers of Russian energy and issue other punishing sanctions against the Kremlin following a pitched battle between lawmakers backing Ukraine and sceptics worried the bill gives the president too much power.

The Lindsey O Graham Sanctioning Russia and Iran Act of 2026 passed in a 262-159 vote on the last day before Congress goes on recess ahead of the midterm elections. Graham, who died earlier this year, co-authored and submitted the bill more than a year and a half ago.

Continue reading...
If you own a Jaguar XJ220 or the rarer XJR-15, you’re probably not short on money or real estate. You may, however, be finding it hard to lay your hands …read more
in charge of / to control (HSK 7-9)

The president of Ecuador is turning the country into a dictatorship by banning parties that don't support him.

The companies that develop Supposed Intelligence (specifically, LLMs) want us to put them in charge of making the systems "safe", by giving them absolute control over what the systems will do (and what they won't do).

The systems are really dangerous, but we cannot trust the companies that develop them to protect us from them. Their interest is to mix more and more subjugation of the public into any protection of the public.

Historically, whenever companies get power over what people's computing facilities can do, they use it against us, building malicious functionalities, designed to benefit them, into the software they invite us to use.