Educational and industry pieces now treat DAO governance as a primary use case for AI. AI agents can summarise proposals, simulate effects on treasury and token price, and predict voting outcomes to help token holders make decisions. Some DAOs experiment with having AI delegates. Instead of humans reading every proposal, you delegate votes to an AI agent with a public track record and guard rails. For example you can place a guard rail that does not allow the agent to vote for proposals that raise protocol risk to certain levels.
Research and investor commentary in 2025 openly speculates about AI centred DAOs. Here AI systems propose, rank and help execute governance decisions. At the end the only role of humans in this system is that of an overseer not day to day voters. However, this system is not all roses, what if an AI swarm quietly steers votes to pass a malicious parameter change. Therefore humans still need to be careful on what power they give to AI agents and to what extent.
The thing is that DAOs are moving from crowd voting to AI assisted decision systems. This brings both a UX win as well as brand new governance risks!
