• Ben's Bites
  • Posts
  • How OpenAI will deal with the first AI elections?

How OpenAI will deal with the first AI elections?

OpenAI is making sure their AI tools like DALL-E and ChatGPT don't mess up elections in 2024. They want the tech to help people safely without interfering with democracy.

What's going on here?

OpenAI outlines what is it doing to safeguard 2024 elections.

What does this mean?

OpenAI has cross-functional teams working hard to prevent their AI from being used to influence elections in sketchy ways. They are red-teaming new systems, building protections, and tweaking policies to reject harmful requests. For example, DALL-E won't generate images of real politicians, and ChatGPT's rules stop people from impersonating candidates or making bots to deter voting.

OpenAI’s updated content policy doesn’t allow ChatGPT use, customs GPTs or making API apps that perform in

  • Political campaigning and lobbying.

  • Chatbots pretending to be candidates or government sources.

  • Create images of real people using DallE

OpenAI is also going to add content credentials to DallE images and they are working on a detection tool as well. ChatGPT will also link to real-time news resources more often to improve transparency.

In addition, OpenAI is partnering with NASS in the US to make ChatGPT point users to official voting info websites when asked about how to vote.

Why should I care?

DALL-E, ChatGPT and other OpenAI tech is catching on fast. How they handle election interference, misinformation, and transparency with these powerful tools will affect all of us. You want AI systems deployed safely and built with strong safeguards.

Join the conversation

or to participate.