OpenAI and Anthropic will each run election "war rooms" for the midterm elections, Semafor reported on Monday. The teams will work to ensure their AI platforms give accurate information about the vote. According to Semafor, neither company has done this before.
An Anthropic spokesperson told Semafor that teams across the company work year-round on election safeguards and on enforcing its policies. The spokesperson said a dedicated war room will monitor the platform for misuse ahead of the election. It will be ready to respond quickly to any incidents. Semafor's report did not include a comment from OpenAI.
Tech platforms first used war rooms in 2018, when Meta set one up over fears of election misinformation. Semafor says AI has made those threats far greater. They now include deepfake videos and AI search results that spread false information. Last week, Google's AI Mode would not answer a question about who won the 2020 election. Google said it was working on a fix.
The work will not be easy, according to Semafor. AI behaves unpredictably, and there have been recent cybersecurity problems. AI search tools also tend to show users the results they want to see. Katie Harbath, a former director of public policy at Meta, told Semafor that war rooms are "built for dealing with the things we can't necessarily predict going wrong." She said such situations need a quick, established process for triaging problems.