The problem isn't the misinformation itself, it's the rate at which misinformation is produced. Generative models lower the barrier to entry so anyone in their living room somewhere can make deepfakes of your favourite politician. The blame isn't on AI for creating misinformation, it's for making the situation worse.