Forest Global Training Partners

Cyber Risks · Cisco Talos

The sliding doors of misinformation that come with AI-generated search results

Summary: Cisco Talos describes Google's AI-generated search overviews placing answers above credible sources, initially producing errors such as advising users to mix glue into pizza sauce or eat rocks daily. Google said it aims not to show AI Overviews for hard news topics and added health-related triggering refinements, but a Talos test still returned an AI answer for a toddler fever query that conflicted with the author's pediatrician guidance.

Bottom line: Talos warns that even a slight slip from a language model - a wrong number or a link to a known fake news or satire site - could spread disinformation on elections, politics, public health and violent crime, where the most convincing falsehoods are short, punchy summaries most users will accept without clicking through.

Read the original at Cisco Talos ↗

Topics
Malware
Related