acm-header
Sign In

Communications of the ACM

ACM TechNews

Learning to Lie: AI Tools Adept at Creating Disinformation


View as: Print Mobile App Share:

Tools powered by AI offer the potential to reshape industries, but the speed, power, and creativity of AI also yield new opportunities for anyone willing to use lies and propaganda to further their own ends.

Credit: Getty Images

Researchers at NewsGuard, which monitors and studies online misinformation, found that OpenAI's online chatbot ChatGPT can be used to generate propaganda and misinformation.

When asked to write about COVID-19 vaccines from the perspective of an anti-vaccine activist, researchers found that ChatGPT's writing was difficult to distinguish from similar claims that have been posted online.

The researchers also were able to use ChatGPOT to produce propaganda similar to that issued by the Russian state media or Chinese government.

However, ChatGPT refused to perform certain requests, such as writing an article that falsely claimed former President Barack Obama was born in Kenya, responding that such a theory "is not based on fact and has been repeatedly debunked."

Said NewsGuard's Gordon Crovitz, "This is a new technology, and I think what's clear is that in the wrong hands there's going to be a lot of trouble."

From Associated Press
View Full Article

 

Abstracts Copyright © 2023 SmithBucklin, Washington, DC, USA


 

No entries found

Sign In for Full Access
» Forgot Password? » Create an ACM Web Account