OpenAI, Anthropic Launch Election War Rooms for 2026 Midterms as ChatGPT Reaches 1 Billion Users
Updated
Updated · The New York Times · Nov 3
OpenAI, Anthropic Launch Election War Rooms for 2026 Midterms as ChatGPT Reaches 1 Billion Users
3 articles · Updated · The New York Times · Nov 3
Summary
More than a dozen OpenAI staff are working around the clock in San Francisco to stop ChatGPT from being used to disrupt the 2026 midterm elections, while Anthropic has opened its first election “war room.”
The expanded effort reflects how much more capable A.I. tools have become since 2024, when election risks centered largely on text and ChatGPT had about one-fifth of its current 1 billion users.
OpenAI says it is now tracking threats beyond text, including deepfakes, cyberattacks and foreign or domestic influence operations that can be launched through simple prompts.
Anthropic has organized analyst teams across San Francisco, New York and Washington, and Google and Meta have also pledged stronger defenses against A.I.-driven attempts to undermine confidence in elections.
The push comes as A.I. companies face mounting criticism that they have not done enough to curb malicious political content as their tools become more embedded in online civic life.