Unless we let it! I always held that the real danger is from giving these algorithms too much power, let them decide things with real world consequences, or hook them up to systems without proper oversight.