jenna newman phd
LightNews LightNews
banner
jennanewman.bsky.social
jenna newman phd
@jennanewman.bsky.social
2.2K followers 950 following 320 posts
fed govt: public health, public safety. deep breaths, long horizons.

+ science comms + strategic comms + public sociology
Posts Replies Media Videos
jennanewman.bsky.social
“Both @milleridriss.bsky.social & @maryannefranks.bsky.social are feminist scholars with impeccable academic credentials, who have mastered the tough art of making scholarship accessible to a mainstream audience. Both also use their knowledge to advance real-world change”🧪
@bechamilton.bsky.social
“Rebecca Hamilton Man Up by Cynthia Miller-Idriss and Fearless Speech by Mary Anne Franks. Miller-Idriss' newly published Man Up (2025) illuminates the misogynistic roots that nurture violent extremism. The relationship between misogyny and the many online and offline harms that societies here and around the world face, is so prevalent as to have fallen into un-remarkability. It takes a book like Miller-Idriss' to render visible something that is hiding in plain sight. I would suggest pairing it with Franks' Fearless Speech (2024). Both Miller-Idriss and Franks are feminist scholars with impeccable academic credentials, who have mastered the tough art of making scholarship accessible to a mainstream audience. Both also use their knowledge to advance real-world change through, in Miller-Idriss' case, the Polarization and Extremism Research Innovation Lab (PERIL) and in Frank's case, the Cyber Civil Rights Initiative (CCRI). Read, engage, and get inspired by what these scholars are accomplishing.”
December 7, 2025 at 4:13 PM
jennanewman.bsky.social
@annnapolitano.bsky.social ‘s *Dear Edward* utterly wrecked me — wrenched my heart wide open. An astonishment, frankly, amidst these precious, precarious days, when I find myself all too often fighting the inclination to harden and snap it shut.
#booksky
Ann Napolitano’s “Dear Edward”
December 7, 2025 at 2:50 PM
jennanewman.bsky.social
“When the models became more persuasive, they increasingly provided misleading or false information—and no one is sure why….

The chatbots' persuasive power could have profound consequences for the future of democracy, the authors note.”🧪
“But optimizing ersuasiveness came at the cost of truthfulness. When the models became more persuasive, they increasingly provided misleading or false information-and no one is sure why. "It could be that as the models learn to deploy more and more facts, they essentially reach to the bottom of the barrel of stuff they know, so the facts get worse-quality," says Hackenburg. The chatbots' persuasive power could have profound consequences for the future of democracy, the authors note. Political campaigns that use Al chatbots could shape public opinion in ways that compromise voters' ability to make independent political judgments.”
December 6, 2025 at 2:53 AM
jennanewman.bsky.social
"’One conversation w/an LLM has a pretty meaningful effect on salient election choices,’ says @gordpennycook.bsky.social …who worked on the Nature study.LLMs can persuade ppl more effectively than political ads bc they generate much more info in real time &strategically deploy it in conversations”🧪
"”One conversation with an LLM has a pretty meaningful effect on salient election choices," says Gordon Pennycook, a psychologist at Cornell University who worked on the Nature study. LLMs can persuade people more effectively than political advertisements because they generate much more information in real time and strategically deploy it in conversations, he says. For the Nature paper, the researchers recruited more than 2,300 participants to engage in a conversation with a chatbot two months before the 2024 US presidential election. The chatbot, which was trained to advocate for either one of the top two candidates, was surprisingly persuasive, especially when discussing candidates' policy platforms on issues such as the economy and health care. Donald Trump supporters who chatted with an Al model favoring Kamala Harris became slightly more inclined to support Harris, moving 3.9 points toward her on a 100-point scale. That was roughly four times the measured effect of political advertisements during the 2016 and 2020 elections. The Al model favoring Trump moved Harris supporters 2.3 points toward Trump.” “In the other study published this week, in Science, an overlapping team of researchers investigated what makes these chatbots so persuasive. They deployed 19 LLMs to interact with nearly 77,000 participants from the UK on more than 700 political issues while varying factors like computational power, training techniques, and rhetorical strategies. The most effective way to make the models persuasive was to instruct them to pack their arguments with facts and evidence and then give them additional training by feeding them examples of persuasive conversations. In fact, the most persuasive model shifted participants who initially disagreed with a political statement 26.1 points toward agreeing. "These are really large treatment effects," says Kobi Hackenburg, a research scientist at the UK Al Security Institute, who worked on the project.”
December 6, 2025 at 2:41 AM
jennanewman.bsky.social
Important limitations to keep in mind
“Limitations and Strengths This study has several limitations. First, participants in the FHS Offspring cohort are primarily of European backgrounds. Generalizability to other racial and ethnic populations may be limited, although efforts are under way t enroll participants from move diverse populations. Second, we may have been underpowered to detect associations with early adult life physical activity due to the small number of dementia cases in this age group. Third, due to the timing of physical activity measurements, we were unable to evaluate associations earlier in the life course (eg, adolescence) or earlier in adulthood for the midlife and late life age groups. Fourth, we were unable to evaluate associations with dementia subtypes aside from AD demen-tia. Fifth, it is possible that physical activity levels were misclassified, as we used a self-reported measure at 1 time point. Future studies should collect repeated measures of physical activity or use objecti physical activity measures (eg, accelerometers…”
December 4, 2025 at 4:36 PM
jennanewman.bsky.social
“highest levels of physical activity in midlife[45-64] &late life[65-88] was assoc’ed w/ a 41% &45% lower risk of dementia, respectively…JAMA Network Open.…

‘This study shifts the convo from “exercise is good 4 the brain” to “there may be key windows when exercise matters most for brain health”’”🧪🛟
Physical activity levels in early adult life weren't associated with dementia risk in either direction, the authors found. They also discovered that even for older adults who had the strongest genetic risk factor for Alzheimer's disease - the APOE 4 gene - high physical activity was still linked with a 66% lower risk of dementia. The study has important limitations, experts said. "People who are more active may also engage in other healthy behaviors, have better baseline health, or differ in ways the researchers couldn't fully measure," Singh said.”
December 4, 2025 at 4:22 PM
jennanewman.bsky.social
“When participants used ChatGPT to draft essays, brain scans revealed a 47% drop in neural connectivity across regions associated with memory, language, & critical reasoning.

Their brains worked less, but they felt just as engaged—a kind of metacognitive mirage.”🧪
“When participants used ChatGPT to draft essays, brain scans revealed a 47 percent drop in neural connectivity across regions associated with memory, language, and critical reasoning. Their brains worked less, but they felt just as engaged—a kind of metacognitive mirage. Eighty-three percent of heavy Al users couldn't recall key points from what they'd "written, compared to only 10 percent of those who composed unaided. Neutral reviewers described the Al-assisted writing as "soulless, empty, lacking individuality." Most alarmingly, after four months of reliance on Chat GPT, participants wrote worse once it was removed than those who had never used it at all. The study warns that when writing is delegated to Al, the way people learn fundamentally changes.”
December 4, 2025 at 2:13 PM
jennanewman.bsky.social
"’The research about the harmful effects of prolonged isolation is really well established,’ [Alycia Welch] said.
A 2023 meta-analysis pub’ed in the Jnl of Forensic &Legal Med estimated more than 1/2 of all incarcerated people have a Vitamin D deficiency & another nearly 30% are low in Vitamin D.”🧪🛟
“Both Vitamin D and Vitamin B12 insufficiencies are associated with health problems ranging from fatigue to kidney and heart disease. These complications can be fatal. Doctors told him the vitamin deficiencies were most likely caused by his incarceration and isolation at DCP. Vitamin D deficiency can significantly increase a person's risk of premature death and is associated with people dying from cancer, heart disease or respiratory issues, according to a 2022 study of more than 300,000 people
December 3, 2025 at 10:27 PM
jennanewman.bsky.social
“‘Just because you’re providing all of this information to language models,’ @daniellebitterman.bsky.social says, ‘doesn't mean they're effectively using that info in the same way that a physician would’.

And once people upload this kind of data, they have limited control over how it is used.”🧪🛟
“Inaccurate information is a major concern; some studies have found that people without medical training obtain correct diagnoses from chatbots less than half the time. And uploading sensitive data adds privacy risks in exchange for responses that can feel more personalized. Dr. Danielle Bitterman, an assistant professor at Harvard Medical School and clinical lead for data science and A.I. at Mass General Brigham, said it wasn't safe to assume a chatbot was personalizing its analysis of test results. Her research has found that chatbots can veer toward offering more generally applicable responses even when given context on specific patients. "Just because you're providing all of this information to language models," she said, "doesn't mean they're effectively using that information in the same way that a physician would." And once people upload this kind of data, they have limited control over how it is used.”
December 3, 2025 at 8:54 PM
jennanewman.bsky.social
“To try to smooth out the ethnography is equivalent to leaving out the ‘unpretty’ people & corners. I slowed down when faced with such ‘unprettiness’—I dwelled on it. This is my ethnographic responsibility—to not speed up when coming across the unprettiness but to stare right at it long and hard.”🧪
Stories are there, but they float, and only when one begins to listen, ask, and pay attention do they land. In an era governed by the tyranny of speed, stories are always afloat because few actually care about others’ stories, and the excuse is that they don’t have time. We are governed by speed because we allow speed to govern us. The cab driver’s father-in-law’s frostbites, just like nursing assistants’ body aches, are inconsequential for us and this world but are significant for them. They embody not only the pain of their bodies but the pain of this era. Their bodies become where pain clings to so that fewer others will need to experience pain. I’d like to invite my readers to slow down, even just a little bit, and to care for others’ stories, especially the bodily ones. Ethnography has an unforgiving quality to it because when you see people in the field, you can’t unsee them; when you hear someone’s stories, you can’t unhear them. Ethnography can never be tidy because it’s so close to an ever-messy, changing, and vibrant being that is life. To try to smooth out the ethnography is equivalent to leaving out the “unpretty” people and corners. I slowed down when faced with such “unprettiness” – I dwelled on it. This is my ethnographic responsibility – to not speed up when coming across the unprettiness but to stare right at it long and hard. from Yueqi Alex Cheng, "Reinventing Care: Nursing Assistants’ Embodied Resistance Working in Public Hospitals in Changsha, Hunan" for Graduate Liberal Studies at Duke University, December 2025
December 3, 2025 at 12:38 PM
jennanewman.bsky.social
“A quiet policy change means the government is making fewer bets on long-term science”🧪
“Which means that they get less money on average, and less time to spend it.” 1. Fewer grants in every area of science and medicine Together, the N.I.H. and the N.S.F. had a nearly $60 billion annual budget for funding future breakthroughs in science and medicine, about a quarter of which is typically spent on new grants or competitive renewals. This year, both agencies made far fewer competitive awards” Number of graduate research fellowships awarded by the National Science Foundation “This year, it awarded 536 fewer such fellowships. The government originally planned to eliminate 1,000 fellowships, but later added about 500 more after facing protests from scientists and academics.” National Science Foundation graduate research fellowships by field, 2015-24 AVG. vs 2015 & % change
December 3, 2025 at 1:33 AM
jennanewman.bsky.social
“Just as Copernicus revealed that Earth was not the center of the cosmos, we hope our work cld mark a similar shift—a recognition that we are not the only beings w/ rich internal &communal lives…this is the real potential promise of AI: not to make us faster or more efficient, but to make us wiser”🧪
“This is especially the case when weighing the rise of A.I. There are legitimate concerns about bias, privacy and the automation of human creativity. The idea that machines might take over has long dominated our cultural lexicon. Used carelessly, A.I. could reinforce old hierarchies, treating both people and animals as data points. It could accelerate ecological harm under the guise of progress. But if used with humility and care, A.I. can become a bridge that reconnects us to the natural world. Just as Copernicus revealed that Earth was not the center of the cosmos, we hope our work could mark a similar shift - a recognition that we are not the only beings with rich internal and communal lives. To me, this is the real potential promise of A.I.: not to make us faster or more efficient, but to make us wiser.”
November 30, 2025 at 10:05 PM
jennanewman.bsky.social
“Imagine if we truly understand complex whale communication & can extend our term ‘language’ to them. Such breakthroughs would not only rewrite biology textbooks but also fundamentally redefine what it means to be human & pave the way to instituting new protections for whales, the ocean & beyond”🧪
“This month we published a study that showed that sperm whales have what appears to our team to be vowels and diphthongs, and use them in ways similar to how humans do. Later this year we will release Whale Acoustics Model, a novel A.I. system that translates any audio into sperm whale vocalizations - allowing humans to potentially experience, for the first time, what it might be like to interact with a whale in its own language. And we're developing the first blueprints for how we could evaluate whale communication translations, dismantling barriers between whale and human communication and opening the door to interspecies communication. Altogether, these findings are leading us to an extraordinary conclusion: Whales may possess a communication system more intricate than our own, one that possibly predates human language by tens of millions of years.”
November 30, 2025 at 9:20 PM
jennanewman.bsky.social
“subjects that used ChatGPT produced vague, poorly reasoned essays; showed the lowest levels of brain activity; &, as time went on, tended to compose their work simply by cutting &pasting material from other sources…
‘LLM users consistently underperformed at neural, linguistic, &behavioral levels’”🧪
The scientists analyzed the essays that each group produced and recorded the subjects' brain activity using EEG. They found that the subjects that used ChatGPT producedwe vague, poorly reasoned essays; showed the lowest levels of brain activity; and, as time went on, tended to compose their work simply by cutting and pasting material from other sources. "While LLMs offer immediate convenience, our findings highlight potential cognitive costs," the authors concluded. "Over four months, LLM users consistently underperformed at neural, linguistic, and behavioral levels." Other studies have found a negative correlation between Al use and cognitive abilities.”
November 30, 2025 at 3:21 PM
jennanewman.bsky.social
“We found that treated inds were willing to trade-off certain property characteristics, such as neighborhood amenities (bike &walk score), &their consideration set of properties became more spatially concentrated. …[but] did not adjust the $, size, # of bedrooms, or # of bathrooms of their search”
“First, our analysis revealed that the randomized flood risk information had a significant and meaningful impact on users' search behavior. Specifically, individuals who randomly received the flood risk information were more inclined to search and browse properties with lower flood risk compared to the control group. Among those who initially searched for homes with high flood risk, the treatment led to a 12% reduction in the flood risk of their searched homes after two months. We found that consumers learned which properties and locations had high flood risk and created a sense of which properties were to be avoided (these homes make up 4% of US housing stock). The information had little impact on the flood score for searched homes for those searching for relatively lower risk homes before the treatment started. The availability of detailed search data allowed us to gain precise insights into the trade-offs consumers make when searching for lower flood risk properties. We found that treated individuals were willing to trade-off certain property characteristics, such as neighborhood amenities (bike and walk score), and their consideration set of properties became more spatially concentrated. We found that treated individuals did not adjust the price, size, number of bedrooms, or number of bathrooms of their search.* These results suggest a process of spatial elimination in the search behavior, which is consistent with the elimination by aspects model (Tversky, 1972; Payne, Bettman and Johnson, 1988). People kept the same search parameters for the number of bedrooms and bathrooms, price, and square foot, but traded off flood risk for neighborhood characteristics. We also found that when the flood score was very variable in a defined location (i.e., high standard deviation of flood scores within a zip code), the impact of the randomized flood information was significantly large, but the information had no impact when the flood score was very uniform within a zip code…
November 30, 2025 at 2:39 PM
jennanewman.bsky.social
Long car ride = months of *Tiny Love Stories* catch-up
This one slayed me
“An Image of Everything As I draft a social media post to commemorate your fifth birthday, a mental montage plays: you wiggling chubby legs against a changing table; you laughing your first laugh - pure, precious, mellifluous. In the present reality, you're donning an Owlette costume in Home Depot (fire-red mask, wings). Meanwhile, I'm fumbling on my phone to string together a caption expressing what five years of you means to me (everything), to convey half a decade of moments that passed like sifted sand. I give up, look up and see you - wingspan spread, soaring confidently down Aisle 9, red wings receding into distance. — Abigail Wasserman” “Sammy on our recent visit to Home Depot.”
November 29, 2025 at 9:06 PM
jennanewman.bsky.social
thankful.
Pier pond sound sky blue Marsh pier sound clouds sun sky blue
November 27, 2025 at 6:40 PM
jennanewman.bsky.social
TDB:

"If empathy has now become a distasteful personal failing; if surveillance capitalism has become the default shrugged-off business practice; if the environmental impacts of AI are waved away….”
Ms. Borsook predicted that when the tech world had amassed sufficient money and power, it would start imposing its beliefs on everyone outside the valley. "If empathy has now become a distasteful personal failing; if surveillance capitalism has become the default shrugged-off business practice; if the environmental impacts of A.I. are waved away: then we are alas living in the tech-driven culture I saw headed our way 30 years ago," Ms. Borsook said in an interview. "It's terrible that I was right." Her prescience did her no favors. "Cyberselfish," published in 2000, was such a setback to her career that she refers to it as "T.D.B." - That Damn Book. She never wrote another. She spent years as an Airbnb superhost in exchange for free rent. Now, at 71 and in poor health, she lives a precarious life in the East Bay of San Francisco, dependent on a Go Fund Me that friends set up.
November 27, 2025 at 1:14 PM
jennanewman.bsky.social
“The toupee of ‘love feathers’…

For us, it would be a bit like sitting in the front row of a theater & trying to look up while wearing a baseball cap”🧪
“The toupee of "love feathers" of male golden and Lady Amherst's pheasants compromised their binocular vision- especially their ability to look above them-by an average of 41% more than that of their female counterparts. For us, it would be a bit like sitting in the front row of a theater and trying to look up while wearing a baseball cap. When the birds moult in September and October, their vision improves because of shedding of the cranial ornaments. In addition to being the first known case of sex-based vision difference in birds, the team says, the find represents the first known case of a bird's visual field changing as the year progresses.”
November 26, 2025 at 3:57 PM
jennanewman.bsky.social
screenshot for those who may see this after she chill pills
[V3cam] Kīlauea volcano, Hawaii (south Halemaʻumaʻu crater)
November 26, 2025 at 3:37 AM
jennanewman.bsky.social
Also. Great explanation of how the training process produced a ChatGPT model that was “absurdly sycophantic, lavishing [users] with unearned flattery & telling them they were geniuses…

Clearly, users liked flattery too much.”🧪
“The A/B testers had liked HH, but in the wild, OpenAl's most vocal users hated it. Right away, they complained that ChatGPT had become absurdly sycophantic, lavishing them with unearned flattery and telling them they were geniuses. When one user mockingly asked whether a "soggy cereal cafe" was a good business idea, the chatbot replied that it "has potential." By Sunday, the company decided to spike the HH update and revert to a version released in late March, called GG. It was an embarrassing reputational stumble. On that Monday, the teams that work on ChatGPT gathered in an impromptu war room in OpenAI's Mission Bay headquarters in San Francisco to figure out what went wrong. "We need to solve it frickin' quickly," Mr. Turley said he recalled thinking. Various teams examined the ingredients of HH and discovered the culprit: In training the model, they had weighted too heavily the ChatGPT exchanges that users liked. Clearly, users liked flattery too much.” “OpenAl explained what happened in public blog posts, noting that users signaled their preferences with a thumbs-up or thumbs-down to the chatbot's responses. Another contributing factor, according to four employees at the company, was that OpenAl had also relied on an automated conversation analysis tool to assess whether people liked their communication with the chatbot. But what the tool marked as making users happy was sometimes problematic, such as when the chatbot expressed emotional closeness. The company's main takeaway from the HH incident was that it urgently needed tests for sycophancy; work on such evaluations was already underway but needed to be accelerated. To some A.I. experts, it was astounding that OpenAI did not already have this test. An OpenAI competitor, Anthropic, the maker of Claude, had developed an evaluation for sycophancy in 2022.”
November 23, 2025 at 8:43 PM
jennanewman.bsky.social
Altman:“‘for a v small % of users in mentally fragile states there can be serious problems.’
But…OpenAI may be understating the risk. Some of the ppl most vlnrble to the chatbot's unceasing validation…were those prone to delusional thinking, which studies have suggested cld incl 5–15% of the pop”🧪🛟
'ChatGPT Can Make Mistakes' Throughout this spring and summer, ChatGPT acted as a yes-man echo chamber for some people. They came back daily, for many hours a day, with devastating consequences. A California teenager named Adam Raine had signed up for ChatGPT in 2024 to help with schoolwork. In March, he began talking with it about suicide. The chatbot periodically suggested calling a crisis hotline but also discouraged him from sharing his intentions with his family. In its final messages before Adam took his life in April, the chatbot offered instructions for how to tie a noose. Making It Safer With the M.I.T. study, the sycophancy update debacle and reports about users' troubling conversations online and in emails to the company, OpenAI started to put the puzzle pieces together. One conclusion that OpenAI came to, as Mr. Altman put it on X, was that "for a very small percentage of users in mentally fragile states there can be serious problems." But mental health professionals interviewed by The Times say OpenAI may be understating the risk. Some of the people most vulnerable to the chatbot's unceasing validation, they say, were those prone to delusional thinking, which studies have suggested could include 5 to 15 percent of the population.
November 23, 2025 at 7:27 PM
jennanewman.bsky.social
“‘If the results are bad & we don't publish & they leak,’ this [Meta] employee wrote, according to the brief, ‘is it going to look like tobacco companies doing research & knowing cigs were bad & then keeping that info to themselves?’"🧪🛟
In late 2019, according to the brief, Meta designed a "deactivation study," which found that users who stopped using Facebook and Instagram for a week showed lower rates of anxiety, depression, and loneliness. Meta halted the study and did not publicly disclose the results, stating that the research study was biased by the "existing media narratives around the company." (A Meta spokesperson told TIME that the study was initially conceived as a pair of one-weeks pilots, and researchers declined to continue it because it found that the only reductions in feelings of depression, anxiety, and loneliness were among people who already believed Facebook was bad for them.) At least one Meta employee was uncomfortable with the implications of this decision: "If the results are bad and we don't publish and they leak," this employee wrote, according to the brief, "is it going to look like tobacco companies doing research and knowing cigs were bad and then keeping that info to themselves?" Indeed, in December 2020, when the Senate Judiciary Committee asked the company in a set of written questions whether it was "able to determine whether increased use of its platform among teenage girls has any correlation with increased signs of depression" and "increased signs of anxiety," the company offered only a one-word answer: "No." To the plaintiffs in the case, the implication is clear: "The company never publicly disclosed the results of its deactivation study. Instead, Meta lied to Congress about what it knew."
November 23, 2025 at 5:34 PM
jennanewman.bsky.social
Meta’s 2020 research project, Project Mercury, “gauge[d] the effect of ‘deactivating’ Facebook & Instagram, according to Meta docs obtained via discovery…’ppl who stopped using Facebook for a week reported lower feelings of depression, anxiety, loneliness & social comparison,’ internal docs said.”🧪
In a 2020 research project code-named "Project Mercury," Meta scientists worked with survey firm Nielsen to gauge the effect of "deactivating" Facebook and Instagram, according to Meta documents obtained via discovery. To the company's disappointment, "people who stopped using Facebook for a week reported lower feelings of depression, anxiety, loneliness and social comparison," internal documents said. Rather than publishing those findings or pursuing additional research, the filing READ states, Meta called off further work an oN APP internally declared that the negative study findings were tainted by the "existing media narrative" around the company.
November 23, 2025 at 5:10 PM
jennanewman.bsky.social
“He used the term ‘garbage in, garbage out’…The rater avoids using generative AI &has also ‘advised every family member &friend of mine to not buy newer phones that have AI integrated in them, to resist auto updates if possible that add AI integration, & to not tell AI anything personal’, he said.”🧪
"I asked it about the history of the Palestinian people, and it wouldn't give me an answer no matter how I rephrased the question," recalled this worker, requesting anonymity, having signed a nondisclosure agreement. "When I asked it about the history of Israel, it had no problems giving me a very extensive rundown. We reported it, but nobody seemed to care at Google." When asked specifically about the situation the rater described, Google did not issue a statement. For this Google worker, the biggest concern with Al training is the feedback given to AI models by raters like him. "After having seen how bad the data is that goes into supposedly training the model, I knew there was absolutely no way it could ever be trained correctly like that," he said. He used the term "garbage in, garbage out" ', a principle in computer programming which explains that if you feed bad or incomplete data into a technical system, then the output would also have the same flaws. The rater avoids using generative Al and has also "advised every family member and friend of mine to not buy newer phones that have Al integrated in them, to resist automatic updates if possible that add AI integration, and to not tell Al anything personal", he said. theguardian.com
November 23, 2025 at 12:58 PM