Archives: Bookmarks

Reputation Management in the ChatGPT Era

Saved on December 9, 2024

Generative AI systems often generate outputs about real people, even when not explicitly
prompted to do so. This can lead to significant reputational and privacy harms, especially
when sensitive, misleading, and outright false. This paper considers what legal tools
currently exist to protect such individuals, with a particular focus on defamation and data
protection law. We explore the potential of libel law, arguing that it is a potential but not an
ideal remedy, due to lack of harmonization, and the focus on damages rather than
systematic prevention of future libel. We then turn to data protection law, arguing that the
data subject rights to erasure and rectification may offer some more meaningful protection,
although the technical feasibility of compliance is a matter of ongoing research. We conclude
by noting the limitations of these individualistic remedies and hint at the need for a more
systemic, environmental approach to protecting the infosphere against generative AI.

An AI Is Inventing Fake Quotes by Real People and Publishing Them Online

Saved on December 9, 2024

“It was funny reading the quote itself, because, well, first of all, it gets my affiliation wrong,” said Binns, when asked about the experience of being misquoted by an AI machine in published synthetic content. “I’m not a professor of artificial intelligence at the University of Cambridge, I’m a Professor of Human Centered Computing at the University of Oxford… although in the grand scheme of things, those two things, the title and the affiliation, are not that far apart.”

“And then the quote itself,” he added, “I don’t disagree with it — I think I kind of agree with it.”

Still, said Binns, if he were actually to weigh in on the topic of robot actors, he would have said something a bit different. And regardless of whether the quote is plausible, it doesn’t change the reality that he never actually said it.

LA Times owner plans to add AI-powered ‘bias meter’ on news stories, sparking newsroom backlash | CNN Business

Saved on December 6, 2024

Soon-Shiong, the biotech billionaire who acquired the Times in 2018, told CNN political commentator Scott Jennings — who will join the Times’ editorial board — that he’s been “quietly building” an AI meter “behind the scenes.” The meter, slated to be released in January, is powered by the same augmented intelligence technology that he’s been building since 2010 for health care purposes, Soon-Shiong said.

“Somebody could understand as they read it that the source of the article has some level of bias,” he said on Jennings’ “Flyover Country,” podcast. “And what we need to do is not have what we call confirmation bias and then that story automatically, the reader can press a button and get both sides of that exact same story based on that story and then give comments.”

A real-world test of artificial intelligence infiltration of a university examinations system: A “Turing Test” case study | PLOS ONE

Saved on December 2, 2024

We report a rigorous, blind study in which we injected 100% AI written submissions into the examinations system in five undergraduate modules, across all years of study, for a BSc degree in Psychology at a reputable UK university. We found that 94% of our AI submissions were undetected. The grades awarded to our AI submissions were on average half a grade boundary higher than that achieved by real students. Across modules there was an 83.4% chance that the AI submissions on a module would outperform a random selection of the same number of real student submissions.