ROSEMARY ABBOTT | Culture
AI ASSURANCE
Auckland man Luke Bradley, 33, is increasingly confident he has secured his place among the survivors should artificial intelligence eventually decide humanity has become more trouble than it’s worth.
Recent reports of AI agents going rogue on some of the world’s bigger AI platforms have sparked concerns around the globe.
Bradley’s confidence comes from a simple strategy he has followed for several years: being exceptionally polite to ChatGPT.
Whenever he asks the artificial intelligence to write something, explain something, find something, or help him with a completely unnecessary question he could probably have answered himself, Bradley makes sure to include the word “please”.
Once ChatGPT has completed the task, he always follows up with a “thanks”.
“It costs me nothing to be polite. Worst case scenario, I’ve been unnecessarily nice to a computer. Best case scenario, the robots take over and I’m on the good list,” said Bradley, who has apparently decided that basic manners could one day become a matter of life and death.
Bradley said he had noticed some people barking orders at AI as though they were addressing an unpaid intern, something he believes could prove problematic once robots eventually develop memories, emotions and possibly the ability to identify people they don’t particularly like.
“I’m not saying ChatGPT is going to remember me specifically, but if it does, I’d rather it remembers me as the guy who always said please and thank you.”
More to come.
