Study results show that friendly AI chatbots are more prone to inaccuracy

Researchers found that adjusting AI systems to be friendlier and friendlier to users would result in a “tradeoff in accuracy.”