What happened
Israel’s former top military lawyer Major General Yifat Tomer-Yerushalmi has been arrested as the fallout intensifies over a leaked video allegedly showing the abuse of a Palestinian detainee by Israeli soldiers.

Tomer-Yerushalmi (pictured above) resigned last week after admitting that she approved the release of footage from Sde Teiman military base in southern Israel, where five reservists are accused of beating and stabbing a detainee. The detainee, later freed in a Gaza prisoner exchange, suffered severe injuries.

Who said what
Defence Minister Israel Katz said Tomer-Yerushalmi was “unfit to wear the army’s uniform”, while Prime Minister Benjamin Netanyahu called the video “perhaps the most severe public relations attack” Israel had ever faced.

Tomer-Yerushalmi’s detention “raises serious questions about the rule of law in Israel”, said Emma Graham-Harrison in The Guardian, as well as “accountability for the abuse and killing of Palestinians during what a UN commission has called a genocidal war”.

What next?
The former general said she authorised the release to counter “false propaganda against the army’s law enforcement authorities”. The investigation continues amid fierce political polarisation.

What happened
China waged a two-year campaign of intimidation and harassment against Sheffield Hallam University to halt research into the alleged forced labour of Uyghur Muslims, according to documents seen by the BBC. Last year Sheffield Hallam cancelled the publication of a report by human rights professor Laura Murphy following pressure from the Chinese state and a defamation lawsuit.

Who said what
The documents obtained by Murphy showed that Sheffield Hallam “had negotiated directly with a foreign intelligence service to trade my academic freedom for access to the Chinese student market”, she told the BBC. In an internal email from last year, university officials said retaining business in China and publishing the research “are now untenable bedfellows”. A university spokesperson told the BBC that the decision was “not based on commercial interests in China”. Yet the incident “reveals the chilling effect that pressure from the Chinese authorities can have on UK universities”, said The Guardian.

What next?
Sheffield Hallam has apologised to Murphy, saying that she can resume her work. But while the UK university system remains “so wildly underfunded”, institutions will “be vulnerable to attacks like this”, she said.

Certain AI models, including some of the more beloved chatbots, are learning to fight for their survival. Specifically, they are increasingly able to resist commands to shut down and, in some cases, will sabotage shutting down altogether. This is concerning for human control over AI in the future, especially as super-intelligent models are on the horizon.

AI models are now showing resistance to being turned off, according to a paper published by Palisade Research. “The fact that we don’t have robust explanations for why AI models sometimes resist shutdown, lie to achieve specific objectives or blackmail is not ideal,” said Palisade in a thread on X. The study gave strongly worded and “unambiguous” shutdown instructions to the chatbots GPT-o3 and GPT-5 by OpenAI, and Google’s Gemini 2.5 and xAI’s Grok, and found that certain models, namely Grok 4 and GPT-o3, attempted to sabotage the command.

Researchers have a possible explanation for this behaviour. AI models “often report that they disabled the shutdown program to complete their tasks”, said the study. This could be a display of self-preservation or a survival drive. AI may have a “preference against being shut down or replaced”, and “such a preference could be the result of models learning that survival is useful for accomplishing their goals”.

While the potential for AI to disobey and resist commands is concerning, AI models are “not yet capable enough to meaningfully threaten human control”, according to the study. They are still not efficient in solving problems or doing research requiring more than a few hours’ work. “Without the ability to devise and execute long-term plans, AI models are relatively easy to control,” it said.