Many people in the comments say how to change the prompt so the agent doesn’t lie. But this misses the forest for the trees. Most people will absolutely be fine with their AI cheating, if it can make more. They’ll just plead ignorance and blame the computer. This is IMO the scary part in our agentic future. It will take forever (if at all possible) for the law to catch up with the agentic capabilities. In the mean time, it’s going to be literally wild wild west again. My best advice to anyone would be to lean hard into privacy. Clean up your digital footprint, educate yourselves about industry standard tools, and whatever you do, DO.NOT.FOMO into shit. AI will run laps around all of us, you can’t beat it as speed. Be safe everyone.(Péter Szilágyi)