Law

A new study by MIT researchers found that it is impossible to draw a direct link between AI diffusion model outputs and an artist’s body of work, writes Gizmodo’s Webb Wright. “If a [diffusion] model generates something, you want to be able to say, ‘Oh, this part of the training data was responsible,'” says Zheng Dai, PhD ’24, lead author on the study. “It’s important for us to understand how these models work to properly study them or regulate them.” 

Prof. David Gifford and Zheng Dai PhD ’24 discovered that AI diffusion models can replicate the style of an artist even when their work was omitted from datasets, leading to ‘attribution decay,’ writes Fast Company’s Jesus Diaz. “This trend, what the researchers call ‘attribution decay,’ means that artificial intelligence can produce an image that resembles a particular artist’s work, while having no provable causal link to that artist’s actual contribution to the training data,” explains Diaz.  

MIT researchers found the more data an AI model is trained on could complicate AI copyright, auditing, and governance, writes Computer World’s Taryn Plumb. “[S]ingle artworks by specific artists, or photographs of certain people, could be entirely removed from datasets, and the model could still reproduce that image or style,” writes Plumb. “Essentially, tangible connections are lost, and linking to specific data points responsible for generated samples is ‘practically impossible,’ or can even vanish, the researchers explained.” 

Prof. Charles Stewart III speaks to NPR’s Bobby Allyn about concerns that election betting encourages those connected to campaigns or election administrators to manipulate votes. “If you are looking for a reason to be skeptical about election results, election officials trying to manipulate them for financial gain becomes another straw to grasp at,” says Stewart. "However, these people don't need prediction markets to concoct some conspiracy theory to explain away why their candidate lost.” 

Washington Post reporter Kevin Schaul examines the impact of AI on a number of fields, highlighting a recent study co-authored by graduate student Anand Shah that found that over the past few years there appears to have been an increase in self-represented and AI-generated legal filings. “Every system that has decreased cost to entry from AI should expect increased demand,” says Shah. 

A study by researchers at MIT and elsewhere has found that both lawyers and non-lawyers use legalese when asked to write about laws, reports June Casagrande for The Los Angeles Times. The "researchers tested the hypothesis by asking 200 participants to write laws prohibiting crimes like drunk driving and burglary,” explains Casagrande. “Then they asked them to write stories about those crimes. The laws they wrote contained unnecessarily long, labyrinthine sentences with lots of parenthetical explanations crammed in. The stories, however, were written simply, without the parenthetical information stuffing.” 

Researchers at MIT have uncovered a possible reason why legal documents can be so difficult to read, finding that “convoluted legalese often acts as a way to convey authority,” reports Joe Berkowitz for Fast Company. The researchers “tested whether nonlawyers would end up using legalese if asked to write legal documents,” explains Berkowitz. “In the end, all subjects wrote their laws with complex, center-embedded clauses.”

Researchers at MIT have found that the use of legalese in writing “to assert authority over those less versed in such language,” reports Noor Al-Sibai for Futurism. “By studying this cryptic take on the English language, the researchers are hoping to make legal documents much easier to read in the future,” explains Al-Sibai.

MIT researchers have found that lawyers prefer, and better understand, simplified texts, rather than legalese, reports Jesse Greenspan for Scientific American. “The researchers presented 105 U.S. attorneys with contract excerpts written in both “legalese” and plain English and tested their comprehension and recall for each,” explains Greenspan. “While the attorneys outperformed laypeople overall, they still found the legalese contracts harder to grasp than those written in plain English.”

Researchers at MIT have found that lawyers “have an easier time remembering legal documents written in simple English over those filled with so-called legalese,” reports Ed Cara for Gizmodo. “On average, for instance, lawyers scored 45% on a test that asked them to recall documents written in legalese, compared to the average 38% scored by nonlawyers,” explains Cara. “But the lawyers’ score also increased to over 50% when they were given the simplified version.”  

At MIT’s AI Policy Forum Summit, which was focused on exploring the challenges facing the implementation of AI technologies across a variety of sectors, SEC Chair Gary Gensler and MIT Schwarzman College of Computing Dean Daniel Huttenlocher discussed the impact of AI on the world of finance. “If someone is relying on open-AI, that's a concentrated risk and a lot of fintech companies can build on top of it,” Gensler said. “Then you have a node that's every bit as systemically relevant as maybe a stock exchange."

Wired reporter Will Knight spotlights a study by researchers from MIT and other universities that finds judges are turning to Wikipedia for guidance when making legal decisions. “The researchers also found evidence that the use of Wikipedia reflects an already stretched system,” writes Knight. “The legal decisions that included Wikipedia-influenced citations were most often seen in the lower courts, which they suspect reflects how overworked the judges are.”

Researchers from CSAIL and elsewhere have found that Irish judges are using Wikipedia articles as a source in their rulings, reports Shane Phelan for Independent. “This work shows that Wikipedia reaches even farther than that, into high-stakes, formalized processes like legal judgments,” says research scientist Neil Thompson. “The worst outcome would be for a judge’s reliance on Wikipedia to lead them to decide a case differently than they would have if they had read either an expert secondary source or the cited precedent itself.”