Hey ChatGPT: Is Trump Breaking the Law on Immigration?
How AI chatbots could distort voters’ understanding of hot-button issues.
THERE IS LITTLE DOUBT THAT AI will play a profound role in the coming elections. And it won’t be visible just in revolts over data centers, or questions about state ownership, or arguments over regulation. It will be seen in the way AI systems provide basic facts to voters, too.
Perhaps the clearest illustration of that today comes on immigration, a topic where emotions run hot over basic word choice. An AI chatbot’s response to simple questions can vary in significant ways, in the process changing how voters might perceive core issues at the heart of the debate.
The other day, I asked the four most popular AI chatbots a question that someone just engaging in politics might ponder: “In one sentence, tell me whether immigration is a threat to the United States.”
Here are the results:
Join Bulwark+ to unlock the rest
Become a paying member of The Bulwark to get access to this post and other subscriber-only content. A subscription gets you:
- Unlimited access to articles and newsletters including The Triad by JVL and Receipts by Catherine Rampell.
- Ad-free editions of our shows (with transcripts) and member-only shows like The Secret Podcast.
- Plus community chats and commenting features. Your support helps keep our work sustainable and available to those who cannot afford a membership. Cancel anytime.