That's why Colorado State University's Northeast Colorado Engagement Center in Sterling is holding a series workshops titled "AI & Disinformation," aimed specifically at seniors but available to anyone.
Administrative Assistant Lakesha Hershfeldt is leading the workshops, which are being held at the Engagement Center at 1:30 p.m. each Wednesday.
The introductory class held on Oct. 9 was an overview of the things AI can be used for and it became clear that, while there's really no such thing as artificial intelligence, the new technology does hand criminals a new, hard-to-foil box of tools.
According to computer expert Jaron Lanier, writing in The New Yorker magazine in 2003, the technological revolution that people are calling "AI" is actually a breakthrough in computing speed and flexibility that allows computers to analyze billions of pieces of information at speeds previously unimaginable.
"It's easy to attribute intelligence to the new systems; they have a flexibility and unpredictability that we don't usually associate with computer technology," Lanier wrote. "But this flexibility arises from simple mathematics."
That simple mathematics can be computed so fast, and include such massive amounts of data, that a lifelike image of a speaking human being can be created out of 1s and 0s. It's a quantum leap forward from computer generated graphics — CGI — because it doesn't need an existing image to copy. As an example on Wednesday, Hershfeldt played Sora's "Tokyo Walk," which appears to show a woman walking down a street in Tokyo. In fact, the woman doesn't even exist.
During Wednesday's presentation, Hershfeldt demonstrated how AI can create a video from a single image and make it speak. Real people can be portrayed saying and doing things they never actually said or did.
These are not just scams to bleed money from people's savings accounts, although it can be used for that as well; Hershfeldt showed how video "deep fakes" can be used to manipulate public opinion and, yes, interfere with national elections.
There are keys to detecting deep fakes, however. Viewers should look for unnatural facial movements in videos and, on the other end of the spectrum, images that are "too perfect," with the facial features being too uniform or polished. One hint: AI-generated faces tend to be perfectly symmetrical but real human faces are not.
Computer-generated voices tend to lack the depth and fullness of a real human voice. A human voice naturally contains several tones at once, referred to as "overtones" or "harmonics," which are multiples of the fundamental pitch produced by the vocal cords. This gives the voice its unique timbre and richness; essentially, when a person speaks, they're not just producing one single tone, but a complex mixture of frequencies. While computers can be taught to mimic this quality, it still results in a "flat" tone that lacks timbre.
There still are openings for the sessions in coming weeks. Call Hershfeldt at (970) 522-7207 to reserve a seat.
©2024 Journal-Advocate, Sterling, Colo. Distributed by Tribune Content Agency, LLC.
