Experts warn AI could become a ‘poisoned apple’

Generative artificial intelligence (AI) could become a “poison apple,” according to a joint report by experts including Tokyo University Professor Fujio Toriumi and Keio University Professor Tatsuhiko Yamamoto.

The report, released Monday, is the second such paper compiled by experts since January 2022.

The 2022 report emphasized the importance of “information integrity”, citing the ability of Internet users to resist misinformation by consuming content from a variety of sources.

The latest report contains issues related to generative AI, warning that generated text can be distributed without knowledge of the source material used to create the content.

Because generative AI tools can generate seemingly plausible responses to queries, experts say the technology could be a ‘poison apple’ that negatively impacts our ‘informational health’. .

The report also mentions the role that telecommunications carriers and advertising agencies should play, and touches on education and IT literacy.

The experts proposed the introduction of a system to measure the “health of information” and the development of IT literacy educational materials.

The 2022 document highlighted issues such as fake news and a “filter bubble” where Internet users encounter only information that reinforces their beliefs.

The latest report has 16 co-authors, including Professor Kuniyoshi Sakai of the University of Tokyo, who specializes in language neuroscience. Mr. Toriumi specializes in computational social science, and Mr. Yamamoto specializes in constitutional law.

Mr. Yamamoto said, “I want to spread the concept of ‘information health’ not only in Japan but also in the world.”