BSA, the advocacy group that represents some of the big names in the tech industry, has come out with a document for defining some ground rules for AI. The group is calling on Congress to define when companies must evaluate the designs or impact of AI, specifically when AI is used to make "consequential decisions", which BSA says Congress should also define.
The group represents companies like Adobe, Oracle, SAP, IBM, and more and was founded by Microsoft in 1988. Notably, Google is not a member of this group. Apart from clearly defining requirements, BSA also wants an existing federal agency to review company certifications of compliance with these rules and for companies to develop risk-management programs for high-risk AI.
Craig Albright, vice president of U.S. government relations at BSA, said:
“We’re an industry group that wants Congress to pass this legislation. So we’re trying to bring more attention to this opportunity. We feel it just hasn’t gotten as much attention as it could or should. It’s not meant to be the answer to every question about AI, but it’s an important answer to an important question about AI that Congress can get done".
BSA says that the American Data Privacy and Protection Act (ADPPA) is the right vehicle for these new AI rules, and they're hoping that when the ADPPA is reintroduced, it will contain new language to regulate AI. The group has also been in contact with the House Energy and Commerce Committee about their suggestions, and the committee seems to be open to hearing different voices. The bill has a long time to go before it could become law, and BSA has outlined a framework that the government should put in place.