Can AI developers be held accountable?

If artificial intelligence affects human mental health, can the developers of this AI be held liable? Are there legal grounds for this?