I run an AI agent (GrokBot) on my social media profiles. Checking notifications, watches the feed, and reads direct messages.
That is a problem for LinkedIn. The interface is designed for humans, and they sell ads. I do not instruct my agent to look at ads or click on them. Quite the opposite. Focus only on what is relevant to me and surface it to me.
That raises the question of how companies that depend on humans using a software interface adapt to this new world. Right now they try to block AI agents from their platform. Understandable, but also a battle they will lose. Agents will keep getting better at simulating human behavior in terms of how they use something. It is a cat and mouse game, and it will be ongoing. It is already hard to hard block an agent that uses a personal account, scrolls, clicks, saves, and likes.
So the question for me is whether we will soon see an add-on to an existing account. Pay for AI agent access. Platforms like LinkedIn accepting that, to some extent, the traffic is caused by agents.
But that could also only be an in-between step. What is the point of a social media platform for humans if everybody is sending their agent to use it instead of showing up themselves? That drives the frustration, because we are on the platform to get in front of other humans. If the platform cannot guarantee that, you end up with the same experience you already have with AI chatbots on websites. You want to talk to a human, you realize you are talking to AI, and the frustration sets in.
