“Google researchers publish paper about how AI is ruining the internet”
“Google: AI Potentially Breaking Reality Is a Feature Not a Bug”
I saw these headlines in Futurism and 404 Media. They’re powerful statements and came from a review of a paper by researchers at Google’s DeepMind titled Generative AI Misuse: A Taxonomy of Tactics and Insights from Real-World Data.
Researchers concluded that “potential for harm is significant” and, in many cases, it’s already causing harm.
The paper looked at 200 examples of AI misuse to spread falsehoods or deceptive narratives. “Manipulation of human likeness and falsification of evidence underlie the most prevalent tactics in real-world cases of misuse,” according to the researchers. “Most of these were deployed with a discernible intent to influence public opinion, enable scam or fraudulent activities, or to generate profit.”
Part of the problem, the researchers say, is because Gen AI is so easy to use. It requires minimal technical expertise to generate content.
“…the mass production of low quality, spam-like and nefarious synthetic content risks increasing people’s scepticism (sic) towards digital information altogether and overloading users with verification tasks,” according to researchers. “If unaddressed, this contamination of publicly accessible data with AI-generated content could potentially impede information retrieval and distort collective understanding of socio-political reality or scientific consensus.”
Sharon Adarlo at Futurism put it more plainly.
“People are using generative AI to make lots of fake content because it’s really good at doing that task, and consequently flooding the internet with AI slop,” she wrote.
What are the most prevalent examples? Here’s what researchers found.
Keep in mind that the research only examined a handful of examples of misuse.
While Gen AI has a multitude of positive uses, the potential for harm is significant. Will it ruin the internet? We’ll see.
Image: Bing image creator, Pixabay + my average Photoshop skills.
