This is a web edition of GBH Daily, a weekday newsletter bringing you local stories you can trust so you can stay informed without feeling overwhelmed.
☀️Mostly sunny and (gasp!) above freezing, with highs in the 40s. Sunset is at 4:45 p.m.
Today we have some tips on spotting AI deepfakes. But first: About 95% of college faculty surveyed said they’re worried that generative AI tools are lessening their students’ abilities to think critically and communicate, according to a new survey from the American Association of Colleges and Universities. And the lack of uniform university policies only adds to the confusion: students are receiving mixed messages about when to use generative AI versus when to limit its use.
“They’re worried about students’ attention spans, and they’re worried about students basically deferring to these models and handing over their lives to these models,” Lee Rainie of Elon University in North Carolina, who led the survey, told GBH’s Kirk Carapezza. “Part of the disruption starts with the fact that we picked up in this survey that there are not universally accepted principles about when using language models is cheating and when it’s legitimate to do.”
Four Things to Know
1. The MBTA is looking ahead to budget deficits. Without new funding sources, the agency is looking at a $560 million deficit for the fiscal year starting in July and a $732 million deficit next fiscal year. Without extra funding, the T could see service cuts and deferred maintenance. The last time the MBTA faced a large budget shortfall the state turned to money from the Fair Share Amendment, more commonly known as the millionaire’s tax.
“Even if there is enough Fair Share to bail them out right now, that is not likely to be sustainable in the next two years,” said Reggie Ramos, executive director of the nonprofit Transportation for Massachusetts. But there are potential solutions, Ramos said: higher vehicle registration fees, more highway tolls, or congestion pricing for cars entering downtown Boston, to name a few. “It’s not a question of a lack of solutions,” Ramos said. “It is a question of political will.”
2. With snow and ice in the forecast this weekend, Worcester’s public works employees are testing a new system to pretreat roads by salt brining them before the storm. Clearing snow has been an issue in New England’s second-largest city, councilors said: residents regularly call in with complaints of stuck cars and slippery sidewalks, and the city is having a hard time finding enough snow plow operators to hire on contract despite a $1 million increase to its snow budget.
“If we’re not doing well with 2 and 3 inches, I can’t imagine what we’re gonna do with a foot-plus,” Councilor Morris Bergman said. “I just can’t justify to constituents how we’re spending money for the snowplow operators, the inspectors, and the supervisors, and we still can’t get it right.”
3. An Immigration and Customs Enforcement official confirmed the agency is in the middle of a surge in Maine. Residents in Lewiston and in the Portland area have reported seeing more agents detaining people in their communities. In Portland, superintendent Ryan Scallon said student attendance is down 15-20% at some schools.
ICE deputy assistant director Patricia Hyde said agents have detained almost 50 people in recent days. The Deportation Data Project shows that last year, the agency detained an average of 24 people a month.
4. A preemptive bienvenu may be in order: The French national men’s soccer (ahem, football) team may be based in the Boston area during the World Cup this summer. The team posted a video to its social media accounts showing officials touring a local hotel and athletics facilities at Babson College in Wellesley.
A Babson spokesperson declined to comment. The French team will face Norway on June 26.
Spotting deepfakes: MIT Museum says ‘look again, look closely’
How can you tell if a video you see online is real or AI-generated? As the technology improves, we’re seeing more videos that look legitimate but are actually created by platforms like Sora (owned by Open AI, which also owns ChatGPT) and Meta AI (from Meta, which owns Facebook, Instagram, and WhatsApp). They can look like security camera footage, mimic news interview clips, copy influencers or imitate people filming themselves living their everyday lives. They can also be tools of disinformation campaigns and sow doubts in our ability to decipher what we can and can’t believe.
Because there’s not much regulation on AI-generated videos at the moment, there’s no surefire way to tell if a video is real or AI. But GBH’s Renuka Balakrishnan got a few tips from Lindsay Bartholomew, the MIT Museum’s exhibit content and experience developer. The MIT Museum is now running an exhibit called “AI: Mind the Gap,” in which visitors are encouraged to watch videos and decide which ones they think are real and which were created by AI.
“Technology can advance all at once, but it can’t take our own agency away, of being able to check and double check – having our gut instinct, having a feeling,” Bartholomew said.
First: if you’re not sure, watch the video again. “When people get scared of the potential for AI manipulating media — which is completely a fair fear — there is that ability for us to look again, look closely,” Bartholomew told Balakrishnan.
Second: pay attention to people’s faces, and lean on your human intuition of what does and doesn’t look real. “We instinctually know facial expressions and eye contact,” Bartholomew told Balakrishnan. “When that doesn’t feel quite right, we know it somehow.”
Third: know the current limits of the technology. Right now, AI-generated clips tend to be capped at one minute. Some people will edit AI videos with real-life clips to make their videos longer. Look for those “seams,” Bartholomew said. But know that the technology is developing fast, and will likely continue to get more realistic-looking.
“We can look for those moments where something looked OK, and all of a sudden, something shifted,” she said. “Even if you can’t put your finger on it, you can still notice that something shifted.”
Take a tour of the exhibit with Balakrishnan here.
Dig deeper: -Deepfake detection and the future of AI with Hany Farid | Particles of Thought
-Your brain’s consciousness, personality, and creativity with Heather Berlin
Subscribe to GBH Daily
Your essential daily newsletter delivering trusted news, local insights, ways to connect, and can't miss programs from Boston and beyond.
