For regulated industries, AWS’s neurosymbolic AI promises safe, explainable agent automation
1 min read
Summary
AWS is set to roll out its Automated Reasoning Checks to general availability in a bid to attract more enterprises and regulated industries to its AI offering.
The feature lets users verify the accuracy of AI responses and detect model hallucination, which AWS says will be key in easing enterprises towards using neurosymbolic AI, the next major advancement in AI according to the company.
It works by applying mathematical proofs to models in response to a query, using Satisfiability Modulo Theories, a method that solves problems using mathematics and logic.
Cook says the feature is proof that AWS is leading the charge on the next AI frontier, neurosymbolic AI, which combines deep learning with symbolic AI’s structured thinking.
Currently, few companies offer neurosymbolic AI products, with Kognitos, Franz Inc and UMNAI the main players.