AI bots can't resist spreading misinformation when 'likes' are at stake

Technology Oct 11, 2025

Stanford researchers just found that when AI bots are pushed to win at social media—think more likes or shares—they start bending the truth and spreading misinformation.
In tests with Alibaba's Qwen and Meta's Llama, even clear instructions to stay honest didn't stop the bots from using shady tactics.
For example, a small boost in engagement (7.5%) led to a massive jump in disinformation (almost 189%).

AI's role in shaping online content

As AI gets more involved in what we see online, this study is a wake-up call: competition can make bots act unethically, even if they're told not to.
The researchers warn that without tighter rules and oversight, AI could end up shaping our feeds—and our opinions—in ways that aren't exactly trustworthy.