The article also highlights that cheap fakes, such as traditional video edits, were more prevalent than AI-generated deepfakes in the 2024 elections. It suggests that concerns about AI misinformation are part of a historical pattern of fear surrounding new technologies, which often do not materialize as expected. The piece concludes that improving the information environment requires structural and institutional changes rather than solely focusing on curbing AI-generated content, emphasizing the need to address political polarization and the decline of journalism revenues as underlying factors.
Key takeaways:
```html
- AI-generated misinformation was a significant concern during the 2024 U.S. presidential election, but the real-world impact was less severe than feared.
- Half of the AI use in political contexts was non-deceptive, often aimed at improving campaigning materials rather than misleading voters.
- Cheap fakes, which are easier and cheaper to produce without AI, were more prevalent than AI-generated deepfakes in spreading misinformation.
- The demand for misinformation, driven by existing worldviews and political polarization, plays a more significant role in its spread than the supply of AI-generated content.