Interview A notable flaw of AI is its habit of

Can AWS really fix AI hallucination? We talk to head of Automated Reasoning Byron Cook

submited by

Style Pass

2025-01-07 13:30:16

Interview A notable flaw of AI is its habit of "hallucinating," making up plausible answers that have no basis in real-world data. AWS is trying to tackle this by introducing Amazon Bedrock Automated Reasoning checks.

Amazon Bedrock is a managed service for generative AI applications and according to AWS CEO Matt Garman, who spoke at the re:Invent conference in Las Vegas last month, the checks "prevent factual errors due to model hallucinations... Bedrock can check that the factual statements made by models are accurate."

These are big claims. What lies behind them? Byron Cook, who leads the AWS Automated Reasoning Group, while also being professor of computer science at University College London, tells The Register:

A lot of people don't understand that they've tricked programmers into doing proofs of memory safety, and the borrow checker in Rust is essentially a deductive theorem prover. It's a reasoning engine ...

"I've worked in the space of formal reasoning and tools for doing that. I brought this sort of capability to Amazon starting about 10 years ago, and then there's been some application to AI. Now suddenly my area, which was extremely obscure before, is suddenly not obscure."