Age gating will require age verification — and that will change the internet as we know it. As I write in @theguardian.com, when users must identify themselves, they may self-censor: thinking twice before sharing opinions or memes fearing retaliation or stigma www.theguardian.com/commentisfre...

At #CPDP2026 in Brussels presenting research on children's safety & rights. I'll share findings from my report with Michal Luria, Ph.D. — centering teens & parents on proposed safety measures that uphold rights, family autonomy & an open internet. cdt.org/insights/wha...

Incredibly important new research articulating the impact of "institutional influence" on LLM outputs. Important to view through the lens of LLM outputs in non-English langs where training data comes from state-sanctioned sources, incl. state media, govt agencies, and even state-selected literature:

New in Nature: LLMs give "the party line" in the languages of authoritarian regimes. This works when they control the media, which feeds pretraining data. We show more state control over the media means less critical LLMs. 6 studies spanning 38 languages & 13 models. Details ↓

"This is not the time to hand over more data about ourselves to tech actors." @aliyabhatia.bsky.social Indeed.

"Want to Google your symptoms, join an ICE watch group on Facebook or scroll Reddit? You might need to show ID," writes @aliyabhatia.bsky.social. "Age verification is coming to platforms near you. Worse, it’ll come at the expense of your rights."

We don't have to imagine an internet without intermediary liability protections like Section 230: in India, a YouTuber's review of a new Motorola phone has resulted in not only a takedown of the video but a civil defamation case against platforms hosting the review: restofworld.org/2026/motorol...

BREAKING: The Chinese government pressured Zambia to screen + exclude Taiwanese speakers attending @rightscon.org, leading to the largest international convention on digital rights being canceled a week before it takes place, the host org @accessnow.org tells WIRED. By @telliotter.bsky.social + me

Following our April 29 announcement, we believe it's important to be transparent about the context that led to the decision. We want to explain, where we can, why this was made on such short notice, only days before we were set to host RightsCon. Read: www.rightscon.org/rc26-stateme...

Happy birthday to #GoogleTranslate which turns 20 today! Google Translate has helped me in countless situations: from translating my grandmother's Gujarati WhatsApp forwards to helping me explain an ankle injury to a pharmacist in Guatemala. It's a tool, it's a game, it's essential.

TIL there is no sickle emoji on iMessage, tho the hammer and sickle emoji was approved by Unicode 1.1 in 1993. Give us the hammer and sickle emoji!! emojipedia.org/hammer-and-s...

Such a gutting day. I feel for the staff of @accessnow.org, all their partners and staff in Africa who have been pushing for a RightsCon to be held on the continent, and for human rights broadly. We're ready to support and show up whenever we get the chance next: www.404media.co/rightscon-hu...

What will it take to create a multilingual and participatory #AI that works for everyone? @aliyabhatia.bsky.social, Jhalak M. Kakkar, Dhanaraj Thakur and ECNL's @marle.bsky.social explain in this new @techpolicypress.bsky.social blog:

Deciding who gets to shape AI is closely related to deciding who benefits from AI. Marlena Wisniak, Jhalak Kakkar, Dhanaraj Thakur, and I articulate what it will take to build a paradigm for participatory, multilingual, and plural AI that truly works for all www.techpolicy.press/a-multilingu...

Also India is now the worst of the South Asian countries on this index.

"The State Department is excluding tech researchers from the United States because of their constitutionally protected work, and this newly disclosed memo only underscores the unbounded scope and unconstitutionality of the policy" -- @carriedecell.bsky.social knightcolumbia.org/content/knig...

AI is often marketed as a "miracle cure" for social challenges, but its success in the developing world depends more on human infrastructure than model sophistication https://restofworld.org/2026/ai-social-good-humans/?utm_campaign=row-social

One corollary of the current local LLM drive is that it puts the power to train and finetune new models into the hands of people outside big tech giants. The result? More representation of global majority cultures and language. My latest for @fastcompany.com www.fastcompany.com/91527477/the...

Always great to talk to @stokel.bsky.social. Developers and researchers around the world are building local LLMs that work in the world's languages. But barriers still remain. How do these approaches inform the big AI labs? That remains to be seen: www.fastcompany.com/91527477/the...

When companies underinvest in T&S in non-EN langs, they're only exacerbating existing vectors of risk posed to the communities speaking those languages. This is laid bare in @untoldmag.org on the coupling of online and offline risks to LGBTQ+ communities in Egypt untoldmag.org/egypt-lgbtq-...

Really cool new challenge from Adaptation Labs: a 2 week open data challenge to address critical gaps in data either in languages other than English and/or under resourced domains www.adaptionlabs.ai/blog/the-unc...

Bluesky tech policy folks: what is today's version of The Stack by Benjamin Bratton? Or is this still the foundational read for analysis on how governance and infrastructure overlap/intertwine: direct.mit.edu/books/monogr...

In India, a growing number of people are using AI chatbots daily, yet these systems aren’t built for diverse linguistic and cultural contexts, which limits their experience. More from @cdt.org's @aliyabhatia.bsky.social → cdt.org/insights/op-...

We didn't deserve this day.

Flock coming to Canada for the 1st time. Some of Flock's uses in the US: "the Atlanta police chief who used Flock’s ALPR [Automated License Plate Recognition] to stalk and harass people, or the Texas cop who used data from 83,000 ALPRs to track a woman suspected of seeking an out-of-state abortion."

There's a compelling Altman as Ravana metaphor here

A Throne for a King has been set up on the mall and then promptly removed. In India, the police would have showed up, people would have been arrested, tweets and instagram posts would have been ordered to be removed. The state of freedom of expression today: thewire.in/rights/is-th...

NEW: "The Park Service has every right to modify [Malcolm X Park], which so badly needs care, but doing so during the summer hurts all D.C. residents," writes @aliyabhatia.bsky.social, who lives near the park, in a new opinion piece. "Yes, the park deserves care. But so do our neighbors."

For @51st.news, wrote about my favourite park in DC: Malcolm X! When I heard about the park closing for the summer, I was distraught! My neighbor (or Park Uncle) was skeptical - he's heard of enough of these beautification efforts in his 41 years near the park 51st.news/opinion-malc...

Resharing this piece from #CODI member @cdt.org on content moderation in the Global South, with perspectives from social platforms, digital activists and others on Maghrebi Arabic dialects, Kiswahili, Tamil and Quechua. 🌏 More from @aliyabhatia.bsky.social → cdt.org/insights/con...

I'm thinking of this especially in the context of India, where a new trans bill has revoked the right of self determination and identity for millions of people despite bestowing the right mere years ago: www.amnesty.org/en/latest/ne...

On Trans Day of Visibility, an essential piece on the right to selfhood and anonymity online, particularly for communities who don't have the same luxury in real life, and how that is threatened by new age verification requirements/laws: www.theverge.com/cs/features/...

This is both a bold and incredibly nuanced position: Wiki pages train AI systems and the utility and efficacy of synthetic data to train systems begs a lot of questions, so it's smart for Wiki to remain a source of high quality human-generated training data particularly in langs other than English

This is an absolute scandal; the United States government is deliberately consolidating all deportation cases of an ethnic minority that the President has called "garbage" in front of a handpicked group of judges and with dramatically accelerated timeline. A lawsuit was filed yesterday against it.

“So you have all these people from the Global South... They use the same vocabulary that whoever’s coming up with these AI detectors is flagging as AI.” @emmaalpern.bsky.social on the people falsely accused of using AI: non-Native English speakers and neurodivergent writers nymag.com/intelligence...

Younger millennials learned digital security the hard way: learning how to dodge viruses while downloading music on Limewire. Maybe Australian teens are now en route to learn lessons their own way as they work around the social media ban: www.techdirt.com/2026/03/20/a...

Hope Americans play this game and once and for all stop thinking the immigration system is anything but some vibes articulated in the form of rules enforced to favor those with means and hurt those without

Decode is doing such essential reporting and analysis about the lived realities of tech users in India. W/ the Pulitzer Centre, they uncover how FRT used in welfare settings misclassifies pregnant women because they can't understand how pregnancy affects faces... www.boomlive.in/decode/ai-fa...

The party of Small Government demands 24 hour video footage of your children to "keep them safe": H/t @isabelalinzer.bsky.social for sharing this shrewd analysis: www.mcsweeneys.net/articles/the...

I estimated how much revenue the State Department and USCIS have received from legal immigrants and US sponsors for benefits that they are now banned from receiving. The answer is over $1 BILLION. That's just from the people who would've been approved this year cato.org/blog/banned-...

interested in questions of tech, labor, and worker futures? on April 10 in Princeton i'm hosting an outstanding lineup of panelists at "the future of (data) work," where speakers will discuss the ethnographic, archival, and participatory action research they've conducted with data + tech workers

Mamdani’s approach to Ramadan - one potent symbolic Iftar each night - is a model for elected officials broadly. We vest our electeds with a fusion of political+cultural power. All of them, at every level, can use of that civic magic judiciously to alter narratives about who is worthy in society.

all of this ai job loss "research," most of which is totally made up by AI companies themselves, is even more useless because it does not consider the broader impacts AI slop and AI porn is having by killing the human internet, social media, discoverability, etc www.404media.co/ai-job-loss-...

Technical looking AI analyses are being weaponized to falsely discredit authentic evidence: www.techpolicy.press/how-ai-conte...

NEW: Three teens allege xAI and Grok were used to generate nude underage images. Photos from homecoming, yearbook, beach outings were turned into CSAM and distributed on Discord and Telegram + some were bartered for other child abuse imagery, a lawsuit alleges. www.washingtonpost.com/technology/2...

Today is seven years since the deadly Islamophobic attack in Christchurch, NZ. The Christchurch Call emerged in the attack's aftermath has been held up as a crucial example of multistakeholderism. Revisiting what I wrote about the future of the Call and similar initiatives: cdt.org/insights/the...

New coverage from @euronews.com on Prof. @geoplace.bsky.social's research, which finds that answers from OpenAI’s ChatGPT favour wealthy, Western countries and sideline much of the Global South. Read more: www.euronews.com/next/2026/01...

1/ As the US continues AI-enabled strikes on Iran, new @brennancenter.org research from @emileayoub.bsky.social and me examines how the military’s investments in the technology have been building to this moment. Congress must step up and reckon with AI’s dangers to life, liberty, and democracy 🧵: