If Cambridge Analytica (and its co-conspirators) ran 100s of thousands of iterations of the same advertisement and recorded reactions for each down to the individual level in order to cause behavior change, how's that different from no-consent medical experimentation? It's not.