ONE PIECE OF NEWS

🤖The rise of the AI-powered disinformation

Microsoft’s researchers announced on Thursday that they found a network of fake, Chinese-controlled social media accounts seeking to influence U.S. voters using AI. The affected platforms were Facebook and X.

Microsoft said the social media accounts were part of a suspected Chinese information operation. The accounts posted about politically divisive topics, such as gun violence, and attacked U.S. political figures and symbols with text and images crafted by AI.

The problem: While Russia tried to influence the previous US Elections with misinformation campaigns, this time the technology is far more capable.

A recent study by the University of South Florida showed that even linguists from top journals could only identify AI-created content 38.9% of the time. The study also found that longer texts are easier to identify as they tend to contain more hallucinations, meaning that with shorter social posts, we have an even lower chance of distinguishing between human- and machine-generated texts.

China seems to be only testing the waters for now, but it remains to be seen when and how it will deploy this new political weapon at scale.

This is a pressing issue for companies and policymakers in western democracies, as an AI-dominated social media landscape would just strengthen echo chambers and undermine public discourse.

As individuals the best thing we can do against AI-generated misinformation is to stay skeptical, engage with people outside social media and to avoid sharing it.

ONE IDEA

If you flood the public sphere with millions and potentially billions of AI agents that can hold conversations and tailor their words specifically for you, it will ruin the conversation between people.

It will ruin the trust between people. You will no longer be able to have a democracy in this situation.

Yuval Noah Harari the best-selling author of ‘Sapiens’ on the Lex Fridman podcast at 00:38:37

⌚ If you have one more minute:

• Microsoft to defend customers on AI copyright challenges

• Facebook Trains Its AI on Your Data. Opting Out May Be Futile

• What AI already does well in supply chain management

• AI-narrated audiobooks are here – and they raise some serious ethical questions

• Google required to stamp out AI-generated images of child abuse in Australia

AI Art of the day 🎨

Scenes from the world of The Last Of Us as miniature tilt-shift photography created with Midjourney by u/Jespergylding. See the whole collection here.

🌊🏄🌊🏄🌊🏄🌊🏄🌊🏄🌊🏄🌊🏄🌊🏄🌊🏄🌊🏄🌊🏄🌊🏄

That’s it folks!

If you liked it, please share this hand-crafted newsletter with a friend and make this writer happy!