The propensity for AI techniques to make errors that people miss has been on full show within the US authorized system as of late. The follies started when attorneys submitted paperwork citing instances that didn’t exist. Related errors quickly unfold to different roles within the courts. Final December, a Stanford professor submitted sworn testimony containing hallucinations and errors in a case about deepfakes, regardless of being an skilled on AI and misinformation himself.
Now, judges are experimenting with generative AI too. Some consider that with the suitable precautions, the expertise can expedite authorized analysis, summarize instances, draft routine orders, and general assist pace up the courtroom system, which is badly backlogged in lots of elements of the US. Are they proper to be so assured in it? Learn the complete story.
—James O’Donnell
What you might have missed about GPT-5
OpenAI’s new GPT-5 mannequin was supposed to provide a glimpse of AI’s latest frontier. It was meant to mark a leap towards the “synthetic normal intelligence” that tech’s evangelists have promised will remodel humanity for the higher.
In opposition to these expectations, the mannequin has principally underwhelmed. However there’s one different factor to take from all this. Amongst different strategies for potential makes use of of its fashions, OpenAI has begun explicitly telling folks to make use of them for well being recommendation. It’s a change in method that alerts the corporate is wading into harmful waters. Learn the complete story.
—James O’Donnell
This story initially appeared in The Algorithm, our weekly e-newsletter on AI. To get tales like this in your inbox first, enroll right here.