Lightnews — Scholar-powered news

sorelle

@friedler.net

2K followers 530 following 47 posts

CS prof at Haverford, Brookings nonres Senior Fellow, former White House OSTP tech policy, co-author AI Bill of Rights, research on AI and society, @facct.bsky.social co-founder formerly @kdphd 🐦 sorelle.friedler.net

Posts Media Videos Starter Packs

Pinned

sorelle @friedler.net · Aug 19

New @brookings.edu post up by me and @aselbst.bsky.social describes 5 points of bipartisan agreement (yes - agreement!) on how to regulate AI... and 2 points of disagreement.

www.brookings.edu/articles/fiv...

It's the listicle version of a forthcoming article: papers.ssrn.com/sol3/papers....

5 points of bipartisan agreement on how to regulate AI | Brookings

Sorelle Friedler and Andrew Selbst detail bipartisan agreements between the Trump and Biden administrations on how to regulate AI.

www.brookings.edu

2 9

Reposted by sorelle

Dan Greene @dmgreene.bsky.social · 1d

Incisive, beautiful reflection from @bedoyausa.bsky.social on his time in the FTC and how it upended his view of the world and his theory of change for it. We really can win, but we don't win by running from the fight.

How I Became a Populist

My time at the Federal Trade Commission—before Donald Trump fired me—totally changed the way I see our political divide.

newrepublic.com

7 21

sorelle @friedler.net · 1d

Tech folks sometimes act like no one else can understand AI or its impacts like we do. That's just not true. Communities know how these systems impact us, and our representatives can act to prevent harm.

Glad to see Philly City Council starting that important work!

#PhillyAIHearing 4/4

sorelle @friedler.net · 1d

AI is like the past. AI is like redlining. AI is like a factory.

I used analogies to help policymakers understand AI better. For the longer version of these analogies, see this piece by me and @maidino.bsky.social

equitablegrowth.org/analogies-fo...

#PhillyAIHearing 3/n

Analogies for AI policymaking

Overview As a suite of technologies under the label of artificial intelligence increasingly take center stage, policymakers recognize the need to develop deep technical expertise about the deployment of evolving technologies in distinct contexts. But specific technical knowledge in the hands of a few will remain insufficient. All policymakers need to develop intuitions and an […]

equitablegrowth.org

1 3

sorelle @friedler.net · 1d

My main point to policymakers is this: AI is not designed to work 100% of the time.

Let's say an AI system has 98% accuracy. 98% is good! But that means it makes errors 2% of the time. The City needs a plan for what to do about those 2%. 2% of our city is a lot of people!

#PhillyAIHearing 2/n

1 3

sorelle @friedler.net · 1d

Philadelphia's City Council held its first AI hearing today in its tech committee chaired by @ruephilacouncil.bsky.social . It's great to see my city taking AI's impacts on all of us seriously!

I was honored to testify on "What is AI?"

As I explained, AI is not magic...

#PhillyAIHearing 1/n

1 1 10

sorelle @friedler.net · 1d

A good rundown of an interesting hearing this morning. #PhillyAIHearing

Kendra Albert @kendraserra.bsky.social · 1d

Councilmember Landau asks about whether the City is using AI for public safety.

CIO says no, they aren't.

Councilmember Landau says that it's a matter of public record that the City is using Shotspotter and facial recognition.

#PhillyAIHearing

Reposted by sorelle

Kendra Albert @kendraserra.bsky.social · 1d

I should probably pick a hashtag, huh? #PhillyAIHearing

Kendra Albert @kendraserra.bsky.social · 1d

A fascinating statement from the head of the Office of Innovation and Technology for the city of Philadelphia that "if anyone says they are an AI expert, they are lying."

...especially bold given how many legitimate AI experts are in the room for the hearing.

3 7

Reposted by sorelle

Kendra Albert @kendraserra.bsky.social · 1d

Fantastic turnout for Philadelphia City Council hearing on AI, spurred significantly by @peoplestechproject.bsky.social.

2 2 18

Reposted by sorelle

Costa Samaras @costasamaras.com · 7d

Picture how big the Hoover Dam is. An absolute unit. The Hoover Dam has a power capacity of 2 gigawatts (GW).

The solar farm that the Admin just cancelled could have produces 6.2 GW of power. That's more than 3 Hoover Dams.

jael holzman @jael.bsky.social · 7d

SCOOP: The Bureau of Land Management says the largest solar project in Nevada — the Esmeralda 7 mega-farm — has been canceled

The news was quietly dropped via a sudden website update with no public word from any of the companies involved or a statement from the agency

@heatmap.news

Esmeralda 7 Solar Project Has Been Canceled, BLM Says

It would have delivered a gargantuan 6.2 gigawatts of power.

heatmap.news

130 2.4K 4.6K

Reposted by sorelle

David Wessel @davidmwessel.bsky.social · 7d

Fact o morning: Three of the six Nobel Prize winners in the sciences are Americans who are immigrants. www.wsj.com/opinion/nobe...

Opinion | Immigrants and American Nobel Prizes

Three of the six U.S. science winners this year are foreign-born.

www.wsj.com

18 48

Reposted by sorelle

Jess Calarco @jessicacalarco.com · 8d

As an expert on help-seeking, I'm not surprised that AI makes it harder for students to develop strong relationships with teachers. Because AI means they can get help without asking an actual human teacher, but it doesn't address the distrust that makes it hard for many students to ask. 1/

Hypervisible @hypervisible.blacksky.app · 8d

“One of the negative consequences AI is having on students is that it is hurting their ability to develop meaningful relationships with teachers, the report finds. Half of the students agree that using AI in class makes them feel less connected to their teachers.”

Rising Use of AI in Schools Comes With Big Downsides for Students

A report by the Center for Democracy and Technology looks at teachers' and students' experiences with the technology.

www.edweek.org

1 22 60

Reposted by sorelle

J. Nathan Matias @natematias.bsky.social · 9d

Genuine question - what are OpenAI’s Sora and other video generation tools good for?

I am honestly trying to understand what is so important that it’s worth the cost. If you have examples, I would be interested to hear them.

futurism.com/artificial-i...

People Are Making Sora 2 Videos of Stephen Hawking Being Horribly Brutalized

People are using OpenAI's Sora 2 to generate videos of theoretical physicist Stephen Hawking being brutalized in ghoulish ways.

futurism.com

15 25 170

Reposted by sorelle

Duaa Eldeib @deldeib.bsky.social · 9d

They are one of the industry’s best-kept secrets, and only a tiny fraction of those eligible actually use them.

How a little-known appeal could force your insurer to pay for lifesaving care

www.propublica.org/article/heal...

This Little-Known Appeal Could Force Your Insurer to Pay for Lifesaving Care. Here’s How to File It.

When a health insurer refuses to pay for your treatment, you may have the right to have the denial reviewed — and potentially overturned — by an independent provider. Here are six steps experts sugges...

www.propublica.org

5 110 290

Reposted by sorelle

Siva @sivav.bsky.social · 9d

I can’t believe my life is being affected by the former CEO of World Wrestling Entertainment

newrepublic.com/article/2013...

Why This Essay Could Cause the University of Virginia to Shut Down

How Linda McMahon’s latest “compact” would do deep and permanent harm to American higher education

newrepublic.com

5 71 250

Reposted by sorelle

The Daily Show @thedailyshow.com · 16d

Excited to introduce Vibes from Meta. Eat your slop, piggies!

110 2.7K 7.4K

Reposted by sorelle

Larry Nemecek — Dr. Trek! @larrynemecek.bsky.social · 17d

“AI is the asbestos we are shoveling into the walls of our society, and our descendants will be digging it out for generations.”

JB_AU @jbau.bsky.social · 18d

This whole section really.

Finally: AI cannot do your job, but an AI salesman can 100% convince your boss to fire you and replace you with an AI that can't do your job, and when the bubble bursts, the money-hemorrhaging "foundation models" will be shut off and we'll lose the AI that can't do your job, and you will be long gone, retrained or retired or "discouraged" and out of the labor market, and no one will do your job.
AI is the asbestos we are shoveling into the walls of our society and our descendants will be digging it out for generations:

10 890 1.9K

Reposted by sorelle

Scott Horton @robertscotthorton.bsky.social · 19d

Larry Ellison envisions a surveillance state in which techbros rule. '“Citizens will be on their best behavior, because we’re constantly recording and reporting everything that is going on,” Ellison said in an hour-long Q&A during Oracle’s Financial Analyst Meeting last week.'

Larry Ellison predicts rise of the modern surveillance state where ‘citizens will be on their best behavior’ | Fortune

Oracle's Larry Ellison believes citizens and police alike will be under constant surveillance of each other.

fortune.com

860 1.5K 2.6K

Reposted by sorelle

Matt Novak @paleofuture.bsky.social · Sep 17

When ChatGPT first took off this was the use case that confused me the most. People said they “only” used it to write “first drafts.” That’s the most important draft! That’s the thinking part!

Oliver Darcy @oliverdarcy.bsky.social · Sep 17

Scoop: Business Insider informed its staff this week that they are allowed to use ChatGPT to generate first drafts of their stories, while also indicating the newsroom will not disclose such A.I. use to readers.

Details in @status.news: www.status.news/p/business-i...

Business Insider and the Bots

The Axel Springer-owned newsroom is buzzing over new ChatGPT writing guidelines—part of an aggressive A.I. strategy pushed by its German parentco and detailed in a memo obtained by Status.

www.status.news

36 330 1.9K

Reposted by sorelle

Jake Grumbach @jakemgrumbach.bsky.social · 26d

Restricting visas doesn’t lead to hiring non-immigrants—it leads to hiring foreigners. For every H-1B visa rejection, multinationals add ~0.4–0.9 foreign employees, especially in R&D hubs like India, China, and Canada.

via @florianederer.bsky.social

How Do Restrictions on High-Skilled
Immigration Affect Offshoring?
Evidence from the H-1B Program
Britta Glennon
WORKING PAPER 27538
DOI 10.3386/w27538
ISSUE DATE July 2020
REVISION DATE February
2023
Highly-skilled workers are not only a crucial and relatively scarce inputs into firms' productive and innovative processes, but are also a critical resource determining competitive advantage. An increasingly high proportion of these workers in the US were born abroad and permitted to work on skilled worker visas. How do multinational firms respond when artificial constraints, namely policies restricting skilled immigration, are placed on their ability to hire scarce human capital? This paper combines visa microdata and comprehensive data on US multinational firm activity to demonstrate that firms respond to restrictions on H-1B immigration by increasing foreign affiliate employment at the intensive and extensive margins, particularly in China, India, and Canada. The most impacted jobs were R&D-intensive ones, but there is some evidence that non-R&D employment was also affected. The paper highlights a means by which firms can circumvent constraining policies and mitigate country-level risk, but it also suggests that, for the average MNC, this means is imperfect; for every visa rejection, they hire 0.4 employees abroad. The most globalized MNCs are the most likely to respond to these restrictions by offshoring, highlighting that firm capabilities—in the form of prior internationalization-shape the decision and ability to offshore in response to skilled immigration restrictions; indeed, these firms hire 0.9 employees abroad for every visa rejection. More broadly, the paper provides evidence of a push factor for internationalizing knowledge activity: artificial constraints on resources result in firms circumventing restrictive policies in ways that may not be anticipated by policy makers.

7 120 340

Reposted by sorelle

Leigh Turner @leighturner.bsky.social · 29d

"The question is whether we will act while AI governance frameworks are still forming or will normalize preventable deaths, discrimination, and erosion of public trust as the cost of doing business." www.science.org/doi/10.1126/... By @alondra.bsky.social

An ELSI for AI: Learning from genetics to govern algorithms

In the United States, the summer of 2025 will be remembered as artificial intelligence’s (AI’s) cruel summer—a season when the unheeded risks and dangers of AI became undeniably clear. Recent months h...

www.science.org

9 13

Reposted by sorelle

Alondra Nelson @alondra.bsky.social · Sep 11

The Human Genome Project's Ethical Legal and Social Implications program dedicated more than 3% of its research budgets to this focus. Imagine 3% of public- and private- sector AI research budgets devoted to its societal impacts.

1 10 33

sorelle @friedler.net · Sep 11

Thanks, AI.

share.inquirer.com/ObEMPw

A screenshot from the linked (gift) article showing energy prices spiking over the summer of 2025 in New Jersey and Pennsylvania.

5 8

sorelle @friedler.net · Sep 4

Yes, I'm definitely interested in this! @grimalkina.bsky.social would be happy to chat. Other folks to check out: @costasamaras.com @strubell.bsky.social @sashamtl.bsky.social

Recent measurement work to look at: dl.acm.org/doi/10.1145/...

Power Hungry Processing: Watts Driving the Cost of AI Deployment? | Proceedings of the 2024 ACM Conference on Fairness, Accountability, and Transparency

dl.acm.org

Reposted by sorelle

Jess Calarco @jessicacalarco.com · Aug 30

Did we not learn nothing from the abject failure of online charter schools?

Washington Post article: "For $65K a year, you can send your kid to an AI-driven private school. There are no teachers, and students study just two hours per day."

23 31 200