In the Trump Era, AI Researchers Receive Amusing Mandate to Deduct ‘Ideological Bias’ from Potent Algorithms
“Under Trump, AI Scientists Are Told to Remove ‘Ideological Bias’ From Powerful Models”
“‘The stakes are high. Unlike other technologies that are primarily destructive, AI is transformative. It gives us the abilities we didn’t have before,’ says Roman Yampolskiy, a computer scientist at the University of Louisville and the author of the paper.”
So, apparently, artificial intelligence (AI) isn’t just going to lead us to our doom. Phew! On the contrary, it’s said to be transformative, which we’re told by Roman Yampolskiy, a computer scientist. Here’s the hope that this doesn’t mean morphing humans into sentient toasters, though.
Yampolskiy wants us to believe that our AI overlords aren’t necessarily coming for us Terminator-style. Instead, they apparently want to provide functions we didn’t have before. Isn’t it generous? Cup holders that regenerate snacks indefinitely, perhaps? Now, that’s a technological revolution everyone can get behind.
But the ground-breaking concept didn’t just emerge out of thin air, peeps. A paper has been published, after all. And no such paper comes to light without the unending efforts of late-night coffees and botched up R scripts. The details seem to revolve around the idea that AI could be allowed to operate in an autonomous way. Sure, because humans do a great job when there are no rules, right? Cue eye roll here.
Impressively, this paper seems to have caught the attention of the AI Safety Institute. An entire institute dedicated to the safety of AI, this is where we’re at. It’s hard not to imagine a room full of people wearing protective gear arguing about whether Siri’s tendency to misunderstand dinner reservations constitutes a safety issue.
Despite the knee-slapping humor potential here, there’s an interesting nugget. The AI Safety Institute has a new directive, and its champion seems to be an exciting “America first” AI strategy. For those perplexed by the oxymoronic statement, join the club. Out of all the things one might consider America first in, “safe AI” isn’t the first to spring to mind. Nevermind though, who needs logic when there’s a flashy directive on the table?
All sarcasm and irony off the table for a moment though, there’s a consensus on the potential implications of AI. Or at least, the echo of uneasy agreement echoes in the tech corridors. Despite the debatable usefulness of the infamous America first slogan in this context, there is a messenger here we can’t afford to kill. The stakes are indeed high. And sarcasm aside, it may be prudent to prepare for the future of AI, however unknown that may be. Fingers crossed it includes those endless snack cup holders.