AI Can be a Risky Proposition

At the risk of sounding like a broken record (Records? Remember them?) we will turn our attention once again to AI. This month, however, it’s more of a cautionary tale.

Inc. Magazine recently reported on a study done by Exploding Topics, a trendspotting company, which found that 92 percent of people don’t check the answers that they receive from AI. That may be nothing new. Output of AI platforms has been suspect almost from day one. And the tales of caveats continue to pour in. There are stories of chatbots hallucinating and outright lying to users, sometimes with disastrous consequences.

A tool company lost customers when its AI customer service bot announced a policy change that didn’t exist. AI output incorrectly claimed that a venture capitalist was a board member of a particular company which it pulled from a blog post that was 100 percent AI fiction. There are reports that lawyers have seen their cases dismissed because AI made up phony cases for them to cite.

Talk about extremes, only 8.5 percent of those surveyed always trust AI overviews but 21 percent never trust them. Trust seems to cut across generations. Across all age groups, more than 40 percent rarely or never click through from AI Overviews to the source material.

Exploding Topics also noted some concerning data. 42.1 Percent of search users reported receiving “inaccurate or misleading content.” Another 35.8 percent found AI Overviews to be “missing important context” while 31.5 percent cited “biased or one-sided answers.” 16.78 Percent of respondents have even experienced unsafe or harmful advice from an AI Overview.

This AI path we’ve started down appears to have as many perils as it does promise. What will be the groundwork for a publisher’s liability due to repeating output from AI that was taken out of context, or information that was misinterpreted or just flat-out falsehood. No doubt it will take some time to sort out the practical and legal implications of it all. Stay tuned.

"A year spent in artificial intelligence is enough to make one believe in God.”

— Alan Perlis, an American computer scientist and professor