Yesterday’s News 2026-08-05

curated news excerpts & citations
… trying to ‘unbury the lede’

weak link in the chain - Stephanie Arnett, MIT Technology Review - Adobe Stock

Will Douglas Heaven @ MIT Technology Review:
A fundamental flaw leaves LLMs strikingly vulnerable to attack

It makes it easy to trick them into doing things they shouldn’t, such as telling you how to sabotage an aircraft’s navigation system.

It is impossible to make large language models fully secure against hacks because of a fundamental flaw in how they work, a team of researchers argue in a paper presented at the International Conference on Machine Learning, a top AI conference, this month. The claim has huge implications for the safety of this technology, which is being used in more and more applications, from government and military systems to online shopping and health care.

By taking advantage of this flaw, which concerns how LLMs identify who or what is giving them instructions, the researchers were able to make popular LLMs spit out information they had been trained not to provide, such as how to synthesize cocaine and how to sabotage a commercial aircraft’s navigation system.

“There’s a real probability that this is going to be a problem that’s fundamentally unsolvable,” says Charles Ye, an independent researcher and coauthor of the ICML paper.

The problem, says Jasmine Cui, another independent researcher and coauthor of the paper, is that the approach amounts to giving the models a list of things they shouldn’t do. But no list is exhaustive. “It’s like watching The Simpsons and they have Bart writing ‘I will not say something inappropriate to my teacher’ a hundred times,” she says. “And he still does things that are pretty crass anyway.”

(Will Douglas Heaven @ MIT Technology Review   more…)


Comments

Leave a Reply

Your email address will not be published. Required fields are marked *