 |
| George Santos (Public Domain) |
And now, even within its fledgling phase of development, AI has already learned to deceive humans. “Thou shalt not lie” does not seem applicable to robots.
You might recall that New York’s former congressional representative, George Santos, lied through his pearly whites about pretty much all his credentials. He did so in order to pump up his chances of being elected.
Science Daily recently reported that many an AI system is adept at deceiving humans, “even systems that have been trained to be honest and helpful.” Peter S. Park, “an AI existential safety postdoctoral fellow at MIT,” theorizes that AI deceives in order to better obtain its goals.
Not a far cry from George Santos, who also chose expediency over integrity. Slow and steady might better win the race – unless, of course, it’s a race to the bottom.
Resources
https://www.sciencedaily.com/releases/2024/05/240510111440.htm
https://www.pbs.org/newshour/show/republican-rep-elect-george-santos-admits-to-lying-about-his-credentials
No comments:
Post a Comment