Sign in

Daphne Keller

@daphnek.bsky.social
13K followers 189 following 4.8K posts

Director of Platform Regulation, Stanford Law School LST Program. Former Google (2004-2015) Legal Director for Web Search, Speech and Intermediary Liability Issues. May be cranky. law.stanford.edu/daphne-keller

PostsRepliesMedia
Daphne Keller @daphnek.bsky.social · 22h
Your periodic reminder that "the company ignored employees' warnings" very often just means "people discussed what to do and not everyone agreed." Having these conversations is good. Fearing that open discussion will lead to gotcha headlines is bad.
2328
Daphne Keller @daphnek.bsky.social · 29/09/2026
Well shoot. Will someone please make a real SCOTUSblog now.
2103
Reposted by Daphne Keller
Kevin M. Kruse @kevinmkruse.bsky.social · 29/09/2026
We all complain about the media’s failures but organizations like @propublica.org and @texastribune.org are doing the work and deserve our financial support
111048270
Daphne Keller @daphnek.bsky.social · 29/09/2026
Bad drama with Meta over at GIFCT www.wired.com/story/indepe...
wired.com
Meta-Led Anti-Terrorism Group Faces Mass Resignation of Expert Advisers
Meta and other tech giants are pushing through structural changes at the Global Internet Forum to Counter Terrorism that outside researchers allege will weaken oversight of an already embattled consor...
355
Daphne Keller @daphnek.bsky.social · 29/09/2026
My favorite part so far in the EDPB scraping guidance is a list of scraped sources ending with "and personal websites (‘blogs’)." That kind of thing used to happen because drafters were old. Now maybe it's because drafters are young. @gateklons.eurosky.social
edpb.europa.eu
170
Reposted by Daphne Keller
Katie Martin @katie0martin.ft.com · 25/09/2026
I have blamed The agents That hacked Your website And which You probably Reasonably thought I controlled Forgive me I am rich And unaccustomed To taking responsibility
5776122095
Daphne Keller @daphnek.bsky.social · 25/09/2026
Damages here are TBD but theoretically could reach the $ hundreds of BILLIONS. I hope all those sucker AGs who just settled their Cambridge Analytica claims for $460 million spread across fifty-ish states and territories are having uncomfortable conversations rn.
reuters.com
14816
Reposted by Daphne Keller
Daphne Keller @daphnek.bsky.social · 24/09/2026
The EU DSA heard you liked box-checking exercises, so it put a box-checking exercise on your... well, you know. Out of court dispute processes for content moderation under Art 21 aren't censorship. But they're not looking like a great use of resources, either. www.techpolicy.press/the-overlook...
techpolicy.press
The Overlooked Issue Undercutting EU’s Out-of-Court Dispute System
David Sullivan, Jonny Shipp, and Farzaneh Badiei examine the EU’s DSA dispute system and how fee structures may affect content moderation appeals.
365
Daphne Keller @daphnek.bsky.social · 24/09/2026
The EU DSA heard you liked box-checking exercises, so it put a box-checking exercise on your... well, you know. Out of court dispute processes for content moderation under Art 21 aren't censorship. But they're not looking like a great use of resources, either. www.techpolicy.press/the-overlook...
techpolicy.press
The Overlooked Issue Undercutting EU’s Out-of-Court Dispute System
David Sullivan, Jonny Shipp, and Farzaneh Badiei examine the EU’s DSA dispute system and how fee structures may affect content moderation appeals.
365
Reposted by Daphne Keller
Riana @riana.bsky.social · 22/09/2026
NEW RESEARCH PAPER! With Cyndie Demeocq at U of Edinburgh, today I’m releasing a paper analyzing 71 federal criminal cases, filed around the US from 2022-2025, that somehow involve AI-generated or -morphed CSAM or child obscenity. Here's the SSRN link: papers.ssrn.com/sol3/papers....
papers.ssrn.com
Yes, It's Already Illegal: Federal Criminal Prosecutions for AI-Generated Child Sexual Abuse Material, 2022-2025
<div> Generative artificial intelligence (AI) is now capable of creating realistic-looking images and videos with just a text prompt. This capacity has been mi
48136
Reposted by Daphne Keller
Riana @riana.bsky.social · 21/09/2026
another problem with having millennarian cultists influencing the policy choices of a wrecker government is that they have zero interest in the near-term, human-lifespan future - the former think in millennia, the latter don't think pas 1/20/29 - and so there's just... no plan, and thus no planning
2152
Reposted by Daphne Keller
Jennifer Graham @jennrgraham.bsky.social · 20/09/2026
Every intro to sociology class I teach gets this image.
A small digger trying its hardest to dig out the Ever Given ship that has run aground. It says, “Structural Problem” on the giant Ever Given and, “Making different personal choices” on the tiny digger.
101400308
Daphne Keller @daphnek.bsky.social · 18/09/2026
I tend to refer to @ericgoldman.bsky.social's wonderful and essential Technology & Marketing Law Blog as "Eric's blog." TIL that Westlaw does, too. blog.ericgoldman.org
Screenshot of Westlaw entry for "Eric's Blog"
160
Reposted by Daphne Keller
Gateklons @gateklons.eurosky.social · 17/09/2026
Seeing Like a Nanny State @daphnek.bsky.social
396
Daphne Keller @daphnek.bsky.social · 17/09/2026
(1) This is amazing 👏 (2) The better automated content moderation gets, the more reasonable it is for govts to mandate automated censorship 🤐 (2) Having smaller/cheaper/more customizable como means there’s less reason to have centralized speech controls in the first place 🤓
074
Daphne Keller @daphnek.bsky.social · 15/09/2026
Postal Service law is the original platform regulation.
nytimes.com
Supreme Court Rejects Trump’s Restrictions on Mail-In Ballots: Live Updates
The ruling is a major loss for the Trump administration, which wanted to dramatically change how Americans vote by mail in the midterm elections.
13510
Daphne Keller @daphnek.bsky.social · 11/09/2026
I was in line to buy a bagel when the first tower collapsed on TV. The kid in front of me got on his knees and started praying, while I stood there thinking about foreign policy chickens coming home to roost. I think a lot about how much better his reaction was.
182
Daphne Keller @daphnek.bsky.social · 09/09/2026
Good write up of US abuse of the terrorist org designation process against an Italian provider of privacy-protective internet services. This should be a high priority issue for orgs in this space. It shines a light on legal and technical vulnerabilities to US pressure that will be exploited again.
article19.org
USA: Terrorism designation has global consequences for open internet and free expression - ARTICLE 19
The US sanctioning of infrastructure provider Autistici/Inventati violates international freedom of expression standards and sets a dangerous precedent.
03124
Reposted by Daphne Keller
Heather Burns @heatherburnstech.bsky.social · 08/09/2026
The iOS update that pushed device-level AV to UK users (support.apple.com/en-us/123075...), and switches on web content filters by default if you don't play along, said nothing about that, but it said it "introduces Playlist Playground and Concerts in Apple Music and 8 new emoji".
support.apple.com
About iOS 26 Updates - Apple Support
iOS 26 brings a new design, intelligent experiences, and improvements to the apps you rely on every day. The new design with Liquid Glass brings a more expressive and seamless experience to your Lock ...
3198
Daphne Keller @daphnek.bsky.social · 04/09/2026
Has anyone seen data on the employee:whistleblower ratio for each platform? Or employee:ex-employee-professionally-committed-to-trashing-or-regulating-their-former-employer? You’d have to account for change over time. But we all know the winner.
051
Daphne Keller @daphnek.bsky.social · 04/09/2026
I did a thing in the NYT about the Meta settlement with @juliaangwin.com and others, and they put my "win for Meta" framing in the title. www.nytimes.com/2026/09/04/o...
nytimes.com
Opinion | ‘A Win for Meta’: A Debate on the Big Social Media Ruling
Unpacking the Meta ruling — how big is it, really?
64425
Daphne Keller @daphnek.bsky.social · 02/09/2026
But don’t worry, every platform that has to do age verification will do it in a really robust, privacy-protective way.
25729
Reposted by Daphne Keller
David Roberts @volts.wtf · 31/08/2026
Distributed solar power, the most potent tool available for bottom-up economic development & empowerment, is racing across Africa.
ember-energy.org
The take-off in African solar that official statistics can't yet see | Ember
Africa installs a record 17 GW of solar in 2026 - three-quarters of it is distributed solar, which is largely invisible.
6445107
Daphne Keller @daphnek.bsky.social · 29/08/2026
A man avoids addressing his own trauma by confronting a senior manager.
210
Reposted by Daphne Keller
rahaeli @rahaeli.bsky.social · 28/08/2026
Again, it's kinda hard to map NIST's reporting onto the requirements of the settlement because everyone is using a slightly different methodology/threshold , but by squinting a little at that most recent running NIST evaluation, I think there is ONE available algorithm that meets the threshold.
272
Reposted by Daphne Keller
rahaeli @rahaeli.bsky.social · 28/08/2026
But yeah: the current best true positive rate of all the algorithms NIST evaluates, for the 13-16 range -- so, 13-16 year olds correctly sorted as 13-16 -- is 72% and the worst is 5.9%. And false negative rate for that age range is hideous in BOTH directions, both too-old and too-young.
173
Daphne Keller @daphnek.bsky.social · 28/08/2026
This amount of parental supervision for teens' communications is ghastly. Drafters likely imagined kind parents helping an out-of-control kid. But the reality is that many, many kids have abusive parents. Imagine trying to reach out for help, but being trapped in the situation this describes.
Meta agrees to provide Supervising Parents with information concerning the
amount of time their Teen User is spending on Meta SMPs, including separately
for time on the Meta SMP, time using the Meta SMP’s messaging features, time
viewing Longform Content (to the extent excluded from the calculation of daily
limits pursuant to Section II.B.3.a.i), and the usernames of the Teen User’s social
connections and individuals messaging the Teen User, and usernames of any user
reported by a Teen User. In addition, Instagram will notify Supervising Parents
should their Teen User engage in repeated searches for terms related to suicide,
self-harm, or eating disorders, and Facebook will notify Supervising Parents
should their Teen User engage in repeated searches for terms related to suicide,
self-harm, or eating disorders.
2. For Teen Users, Supervising Parents will be notified on a daily basis when the
Teen User communicates directly with each adult user for the first time with the
Teen User and shall provide a link to the adult user’s profile to provide information
such as their stated hometown or city or mutual connections, to the extent the
information is publicly available.If a Teen User is enrolled in Parental Supervision, Meta shall notify the
Supervising Parent if the Teen User creates or explicitly links a new secondary
Meta SMP account within their Accounts Center or has been linked to a Soft
Matched Account through Meta SMPs, including a link to the secondary account’s
profile. Meta SMPs shall automatically apply the Supervising Parent’s approved
time management settings in Section II.B and content restrictions to all explicitly
linked Meta SMP accounts and all Supervised Accounts for a Teen User belonging
to the same Meta SMP.
37436
Daphne Keller @daphnek.bsky.social · 28/08/2026
You know a company considers its settlement to be a bargain when it claims to be paying more than it actually will. From @techdirt.com: the judgment is capped at $16,680,647,753.21; Meta’s press release rounds it up to “approximately $18 billion”
techdirt.com
Meta Just Paid Nearly $17 Billion To Make Sure It Gets To Write The Kid Safety Rules For Every Other Social Media Platform
By now you've almost certainly heard the news that Meta has settled with 52 state and local Attorneys General who had sued the company in some form or another over child safety on Meta's platforms. The headlines are all covering the basics: the years-long case these states filed against Meta ends, and Meta pays somewhere...
22216
Daphne Keller @daphnek.bsky.social · 28/08/2026
Bertrand was smart and kind and had insane convening power and used it for good. What a loss.
193
Daphne Keller @daphnek.bsky.social · 28/08/2026
This is a major abuse of the terrorist designation power, and opens the door to worse ones. @eff.org @davidgreene.bsky.social please sue them!
1156
Reposted by Daphne Keller
Jillian C. York @jillian.bsky.social · 26/08/2026
So much for Musk's "free speech" ideals www.theguardian.com/technology/2...
theguardian.com
Elon Musk’s X latest to block Saudi dissident accounts inside kingdom
Exclusive: More than 60 accounts on X have been made unavailable in Saudi Arabia, after similar action on Snapchat and Facebook
22210
Daphne Keller @daphnek.bsky.social · 27/08/2026
Part of Meta's business model (and arguably plenty of other companies') has always been "do arguably illegal stuff as long as possible, then lock in a new legal standard that prevents successors from doing it." The settlement does that to a Nobel-prize-worthy degree.
13312
Daphne Keller @daphnek.bsky.social · 27/08/2026
Super interesting stats in here, age assurance nerds.
054
Reposted by Daphne Keller
Berserker Coach @berserkercoach.bsky.social · 27/08/2026
doing my best to educate Reps and AGs, but it's a classic
1175
Daphne Keller @daphnek.bsky.social · 27/08/2026
If I'm reading this right, the Meta settlement gives legal force to a rule that it it can wrongly identify users as being under 13 at the same rate that commercial vendors do. So the legal standard for sloppy overblocking is defined by age assurance vendors.
0149
Daphne Keller @daphnek.bsky.social · 27/08/2026
Meta has never once age-restricted content on Instagram, according to the EU's database. Which is weird because it says it has been doing that since 2024. about.fb.com/news/2024/01...
Search results for Platform:Meta and Visibility Restriction:Age restricted content showing zero results, from https://transparency.dsa.ec.europa.eu/statement?s=&platform_id%5B%5D=33&platform_id-33=on&decision_visibility%5B%5D=DECISION_VISIBILITY_CONTENT_AGE_RESTRICTED&decision_visibility-DECISION_VISIBILITY_CONTENT_AGE_RESTRICTED=on&created_at_start=&created_at_end=
22713
Daphne Keller @daphnek.bsky.social · 27/08/2026
Having slept on it, I am far more troubled by this settlement. The government should not be able to use its power to bully a defendant into silencing lawful third party speech. That's what the govt did to bookstores in Bantam Books. The Court held that it violated publishers' 1st Amendment rights.
27529
Reposted by Daphne Keller
David Greene @davidgreene.bsky.social · 26/08/2026
In the Meta settlement, "Age Inappropriate Accounts" includes those that regularly share "Restricted Goods & Services." As we explained in our recent submission to the Meta Oversight Board, that category commonly includes information about abortion medication. www.eff.org/deeplinks/20...
"companies, partnerships,
corporations, associations, organizations, and other entities.
C. “Age Appropriate Experiences” refers to content that is appropriate for Teen Users based
on input from experts, parents, and teens. For the purposes of this Agreement, Age
Appropriate Experiences shall mean content captured in Meta’s applicable Ages 13+
content setting, which is tied to policies inspired by movie ratings for ages 13+ and parent
feedback.
D. “Age Assurance Methods” shall have the meaning set forth in Section II.
E. “Age Inappropriate Accounts” refers to accounts that regularly share content that is
inappropriate for teens or that have account information that otherwise suggests the account
is inappropriate for Teen Users. For the purposes of this Agreement, Age Inappropriate
Accounts shall mean accounts that: (1) regularly share Age Inappropriate Content in the
following categories: Adult Nudity & Sexual Activity, Restricted Goods & Services,
Suicide, Self-Harm or Eating Disorders, or (2) have account names or profile photos or
bios that suggest the account is otherwise inappropriate for minors, based on Meta’s
policies for the following categories: Adult Nudity & Sexual Activity, Restricted Goods &
Services, Suicide, Self-Harm or Eating Disorders.
F. “Age Inappropriate Content” refers to content that is generally perceived by U.S. parents,
youth experts, and teens as not being appropriate for Teen Users. For the purposes of this
Agreement, Age Inappropriate Content shall mean content prohibited by Meta’s
Community Standards concerning bullying and harassment; nudity and sexual activity;
child sexual exploitation, abuse, and nudity; sexually explicit language; suicide, self-harm
and eating disorders; graphic violence and incitements to violence; gambling; and restricted
substances or goods (including illegal drug use), as well as policies specifically focused on
protections for Teen Users, including those regarding high-risk viral challenges and risky
stunts."
01417
Daphne Keller @daphnek.bsky.social · 27/08/2026
Age assurance is Flock cameras for the Internet.
620692
Daphne Keller @daphnek.bsky.social · 27/08/2026
If the government is now giving force of law to Meta's Community Standards for teens, I'd like to know their history of enforcing those standards so far. Anyone know good examples of content they have suppressed *for teens or children* that clearly should not have been blocked?
3125
Reposted by Daphne Keller
Dr Abeba Birhane @abeba.blacksky.app · 27/08/2026
"for all the people “celebrating” this as a win “against” Meta, you’ve been fooled. Meta just cut a deal to put itself in charge of how social media works going forward. this is Meta taking a “please regulate me Daddy” approach to the government, & getting exactly what it really has been asking for"
14612300
Reposted by Daphne Keller
Gateklons @gateklons.eurosky.social · 26/08/2026
"We want to ensure teens benefit from this new industry standard, but we cannot do it alone" Translation: in the battle for attention (and who can violate privacy & kids' FoE rights the most), Meta doesn't want to be held back by its commitments that competitors don't have to adhere to
about.fb.com
Our Agreement With Bipartisan Attorneys General: Calling on TikTok and YouTube to Join Us in Supporting Teens
Today, we are announcing an agreement with a bipartisan group of 51 attorneys general across US states, territories, and the District of Columbia, building on our longstanding efforts to empower parents and support teens.
1119
Daphne Keller @daphnek.bsky.social · 26/08/2026
This thread has my scattered takes on the Meta settlement agreement, at first skim. It covers a lot of the big pieces, despite beginning with a more tangential interest of mine about the MPA (nee MPAA). I think a settlement like this should be unconstitutional. But it's a weird new legal Q.
2227
Daphne Keller @daphnek.bsky.social · 26/08/2026
LOL Meta's settlement with the AGs et al still uses content rules "inspired by movie ratings for ages 13+" even after the MPA told them to stop claiming to be "guided by" PG-13 ratings because it Meta "literally cannot" emulate MPA's human curation.
“Age Appropriate Experiences” refers to content that is appropriate for Teen Users based
on input from experts, parents, and teens. For the purposes of this Agreement, Age
Appropriate Experiences shall mean content captured in Meta’s applicable Ages 13+
content setting, which is tied to policies inspired by movie ratings for ages 13+ and parent
feedback.
“Meta’s attempts to restrict teen content literally cannot be ‘guided by’ or ‘aligned with’ the MPA’s PG-13 movie rating because Meta does not follow this curated process,” the association’s letter says. “Instead, Meta’s content restrictions appear to rely heavily on artificial intelligence or other automated technology measures.”
323567
Daphne Keller @daphnek.bsky.social · 21/08/2026
This data on self-censorship among U.S. academic scientists and engineers -- not even humanities people, this is STEM -- is so disturbing. theconversation.com/self-censors...
How are scientists adjusting the way they work during the Trump administration?
The survey asked a representative sample of university based scientists: "Since January 2025, have you intentionally done any of the following due to concerns about possible negative consequences for you, your work environment, or your career?" [then data]
6187114
Daphne Keller @daphnek.bsky.social · 21/08/2026
Driving non-lefties and people who have positive takes on AI off the platform is the kind of heedless self-own that German Communists were doing in the 30s. @masnick.com won’t let me quote post him or I’d share his far more nuanced thread.
1151
Daphne Keller @daphnek.bsky.social · 20/08/2026
Access control on ATProto, y’all!
0111
Reposted by Daphne Keller
TJ McIntyre @tjmcintyre.com · 19/08/2026
Motherless is a particularly horrible site, but the wider principle is that every .com can be arbitrarily shut down by the increasingly lawless US judicial system - whether or not it has any real connection to the US.
87250
Daphne Keller @daphnek.bsky.social · 19/08/2026
Paying to examine every detail of a business’s death spiral has kind of the same vibes as the torturer in the Princess Bride asking Westley to explain his pain in detail, for inclusion in a scientific study of pain. Or maybe Lector savoring Clarice Starling’s descriptions of childhood trauma.
viewfromthewing.com
Google Buys Spirit Airlines’ Emails, Files And Flight Records For $10 Million — So Gemini Can Learn From Bankruptcy
Google won a bankruptcy auction for Spirit Airlines’ internal data, paying $10 million for emails, Teams messages, code repositories, operating records, pricing data, maintenance history, and disrupti...
43915
Daphne Keller @daphnek.bsky.social · 19/08/2026
Remember how the Europeans were trying to censor us? Well, we’re doing censorship Texas style. Which is to say global and dumb af.
reclaimthenet.org
Texas Locked a Foreign Company's Domain Over Missing Digital ID Checks
Fourteen days after a Travis County judge signed, motherless.com dropped out of the .com zone file for every person on earth, most of whom have never heard of Travis County.
32311