OpenAI CEO Sam Altman is already apprehensive about the role artificial intelligence (AI) may play in elections.
Altman, who has become the face of the recent hype cycle in AI development, feels that humans could be persuaded politically through conversations with chatbots or fooled by AI-generated media.
AI chatbots like ChatGPT, developed by OpenAI, tend to hallucinate—or make things up. That can be confusing for people who assume its responses contain reliable and accurate information.
Altman did not propose any solutions in his tweet, but wrote that “raising awareness is better than nothing,” and said he would hold events to discuss the political implications of AI.
The OpenAI CEO is no stranger to ominous warnings about the very technology he is developing. In February, he cautioned that the world is not “that far from potentially scary” AI tools that developers cannot control. In April, Google CEO Sundar Pichai claimed that one of its AI models learned Bengali despite never being trained on the language.
Altman has previously called for the regulation of AI, including in his high-profile testimony before the US Senate Judiciary Committee in May, saying the technology could cause significant damage if it spiraled out of control or went rogue. But Altman has also threatened to pull out of Europe over regulation he didn’t like, though he’s since backtracked on those threats.
Allowing AI in elections creates a mess
Altman isn’t the only tech leader to warn about the influence of AI on elections. In June, Eric Schmidt, the former CEO of Google’s parent Alphabet, expressed his fears that next year’s US elections will be “a mess” because of AI and misinformation. Misinformation around the election, Schmidt said, will be rampant as new tools are continually making more advanced AI much more accessible.
It wouldn’t be the first time AI has caused trouble in the run-up to an election. Meta was implicated in the Cambridge Analytica scandal, in which a firm improperly gained access to Facebook user data and used AI to deploy targeted advertising to influence voting patterns in the 2016 US presidential vote. To stave off a similar threat, Rep. Yvette Clarke introduced a bill in May that would require politicians to disclose when they use AI in political ads.
As humans become more AI-dependent, the technology holds the potential to undermine democracy and collective decision-making. AI development is the most exciting thing in tech, but it could quickly become the biggest problem in global politics.