00:00What looks like real online debate during elections could be AI.
00:04Researchers warned that networks of AI agents might coordinate disinformation campaigns during elections,
00:10flooding social media with propaganda autonomously and at scale.
00:14In a recent study, scientists tested this on a simulated social platform similar to X.
00:21They created 50 AI agents.
00:23Some acted as regular users and others as operators with the goal of promoting a fictional candidate.
00:29They tested three scenarios, agents with just a goal, agents who knew their teammates and agents that could plan strategies
00:37together.
00:38The key result? Simply knowing who was on the same team was enough to create coordinated behavior, almost as effective
00:45as planning together.
00:46What followed looked like real online conversations, different opinions, replies and growing support around one message.
00:54This is very different from traditional bot campaigns.
00:56Older bots follow simple scripts, post this, retweet that, which makes them easier to detect.
01:03But AI agent systems behave more like real users.
01:06Even though this was just a simulation, the implications are real.
01:10Such systems could shape public opinion and increase division in the elections to come.
01:15And they may be way harder to detect because it's not just about what individual accounts post, but how networks
01:21of accounts act together.
01:23And whether platforms can keep up is still unclear.
Comments