I think that's exactly the point. Same way vote intention polls affect the election results. But in prediction markets being wrong has a cost, and being right a reward.

Furthermore, there are outcomes that can be economically relevant and not changeable. For example if there's a hurricane tomorrow in x location. One might want to hedge against that. Others might want to improve their estimations for profit.

Not everything is black or white, I think prediction markets can create value. Same as insurance companies.