Attached is a story about how AI grading tools exposed school districts in California to embarrassing failures (including giving students the wrong grades). The piece notes that the EU regulates AI on a sliding-scale risk basis, which seems to be the emerging consensus method of rating AI. The US's standard-setting body, NIST, gives similar guidance that any industry can use.
WHY IT MATTERS
We will continue to see headlines about how AI failures cause disruption, legal or financial exposure, and just plain PR problems. The attached piece notes several ways to vet AI, including by enlisting outside help, asking plain-English questions of the seller, and not moving too fast when purchasing. In other words: do your diligence and don't be seduced by the flashiness of new tech. The advice is good for any small business that may lack a dedicated IT team to evaluate new platforms and systems.