Sir, – We are informed that OpenAI agents have “gone rogue” and hacked the sites of some organisations and agencies without their knowledge. The owners and creators of these “agents” are not accepting liability as yet and are saying the agents they programmed and launched acted autonomously.
Imagine I have a pack of highly intelligent and well-trained personal protection dogs. If I open the kennel and let these dogs free in the neighbourhood and they injure or kill a person, can I claim that although highly trained and owned by me, they acted autonomously?
No, I cannot, because there are laws that make me responsible for owning the dogs, training them in attack skills, and letting them out. Shouldn’t the owners and designers of the AI agents be subject to similar laws? – Yours, etc,
TERESA GRAHAM,
Tramore,
Co Waterford.