This is why AI safety is going to be impossible. This easily could have been a bad actor who would use this finding for nefarious acts. A person can just lie an

This is why AI safety is going to be impossible. This easily could have been a bad actor who would use this finding for nefarious acts. A person can just lie and there really isn't any safety finetuning that would let it separate the two intents.