Blackburn: Need Rules to Ensure Training for AI Large Language Models Isn't Biased Against Conservatives

On Thursday’s broadcast of Newsmax TV’s “Wake Up America,” Sen. Marsha Blackburn (R-TN) argued that there have to be AI regulations to eliminate bias “from these training mechanisms for large language models so that conservatives are protected.”

Blackburn said, “You need those basic privacy protections for the virtual space, one set of rules for the entire Internet ecosystem, with one regulator, so that people are protected. And, as you well know, protecting our children, the Kids Online Safety Act, protecting our intellectual property holders and our creators, which is the NO FAKES Act, making certain that we’re protecting consumers, making certain that we have ratepayer protection when it comes to utility cost, and also making certain that bias is eliminated from these training mechanisms for large language models so that conservatives are protected. If you put garbage in when you are training these large language models, you will get garbage out. And we need to make certain that these things are not biased.”

She added that consumer protections are needed and regulation needs to be light.

Follow Ian Hanchett on Twitter @IanHanchett