AI and democracy

Peter Coy, Can A.I. and Democracy Fix Each Other? NYTimes, Apr. 5, 2023.

Attitudes about A.I. are polarized, with some focusing on its promise to amplify human potential and others dwelling on what could go wrong (and what has already gone wrong). We need to find a way out of the impasse, and leaving it to the tech bros isn’t the answer. Democracy — giving everyone a voice on policy — is clearly the way to go.

Democracy can be taken hostage by partisans, though. That’s where artificial intelligence has a role to play. It can make democracy work better by surfacing ideas from everyone, not just the loudest. It can find surprising points of agreement among seeming antagonists and summarize and digest public opinion in a way that’s useful to government officials. Assisting democracy is a more socially valuable function for large language models than, say, writing commercials for Spam in iambic pentameter.

The goal, according to the people I spoke to, is to make A.I. part of the solution, not just part of the problem. Noveck, whose lab uses information technology to assist democracy, told me she opposes Elon Musk’s and other tech leaders’ proposed six-month moratorium on the development of A.I. systems that are more powerful than the GPT-4 chatbot. “We should be doubling down on efforts to use A.I. for social good and strengthening democracy,” she said.

Noveck’s lab has used artificial intelligence tools — including one from Iceland called Your Priorities — to help the people of Oakland, Calif., come to agreement on building tiny houses for the homeless, among other projects. She wrote by email, “Ultimately, we can use A.I. to regulate, make laws, govern and to develop institutions that are flexible enough to address the challenges of A.I.”

Likewise, the Collective Intelligence Project argues that “collective flourishing” requires progress, safety and participation of the public in artificial intelligence.

There's more at the link.