✒️@MonaElswah and I dug into the Oversight Board's Advisory Opinion on Shaheed. We write that without high quality benchmarks to evaluate automated moderation tools across langs, Meta risks continuing to burden Arabic speakers with inconsistent moderation: https://t.co/s8Ji2DRhEZ
These Working Groups are based on the three critical pillars of Safe and Trusted AI, Global AI Governance, and Context-driven Infrastructure.
The three Working Groups will be co-led by two organisations representing and bringing in perspectives from around the world to better inform our Global Policy Agenda.
Konstantinos Komaitis, @AtlanticCouncil's Democracy & Tech Initiative, and @urvashi_aneja, @DigiFuturesLab /Global South Network for Trustworthy AI, will co-lead the Safe and Trusted AI Working Group; @JaquePigatto, Data Privacy Brasil, and @MPaz_online, Global Partners Digital, will spearhead the Working Group on Global AI Governance; and @priachetty, Research ICT Africa, and @AliyaBhatia, Center for Democracy and Technology, will lead the Context-driven Infrastructure Working Group.
We are happy to have our co-chairs on board with us in the next leg of our initiative.
@being_on_line User data is an insufficient proxy to measure whether "enshittification" is real though, especially given network effects, lack of data portability, how entrenched these services are especially in areas of low trust in other info sources (often by design e.g. Free Basics)
We’ve built a one-stop flood resource hub with everything you need right now: verified donation channels for flood relief, district-by-district emergency contacts, hospital & blood-bank numbers, live river monitoring, and how to report a missing person.
https://t.co/CB0rHMe5g2
Please share it, bookmark it. Let us know how we can improve this.
(Concept and development by @dibbeshworr)
speaking of user tools to control online experiences, why can't you see all TV and/or movie titles alphabetically on streaming platforms and instead only navigate between choices that are algorithmically recommended and ordered for you?
I call on New York City Mayor Zohran Mamdani to cancel the Madison Square Garden event featuring Mohan Bhagwat, one of the most prominent and dangerous leaders of far-right Hindu fundamentalism in India.
Bhagwat is the so-called supreme leader of the Rashtriya Swayamsevak Sangh (RSS), which was founded in 1925, and is the hardline, ideological backbone of Indian Prime Minister Narendra Modi’s far-right Bharatiya Janata Party. From its early days, the RSS modeled itself after Nazi Germany and continues to base itself on fascist ideas, violence against Muslims and other minorities, and oppression of working people.
Since Modi came to power in 2014, India has seen horrific escalation of mass repression and violence in regions such as Kashmir, and increasing attacks on Muslims, oppressed-caste communities, and working and poor people as a whole. The BJP-RSS leaders and the billionaire class in India have built close links with the genocidal and apartheid state of Israel. Israeli Pegasus spyware has been used to carry out shocking levels of surveillance of Indian activists, public officials and independent journalists.
By refusing to use his Mayoral position to cancel the event and condemn Bhagwat and the RSS, Mamdani is failing to stand with Muslims and other oppressed communities in India and with the Palestinian people. Modi's right wing regime has gone even further down the road of authoritarianism than Trump, and it is the responsibility of socialists to take a stand against it.
During my decade as the lone socialist on the Seattle City Council, I used my office to build movements to win the nation’s first resolution condemning anti-Muslim, anti-poor citizenship laws by Modi and BJP-RSS. I also won the first and only ban on caste-based discrimination outside South Asia. The Modi regime is retaliating against me by refusing to grant me a visa to visit my mother and family in India.
I thank Hindus for Human Rights @Hindus4HR, the Indian American Muslim Council @IAMCouncil, and other organizations who are demanding that Mamdani shut down Bhagwat’s event.
One month after Indian police deployed unlawful and deadly force against peaceful protesters at the Cockroach Janata Party (CJP)-led “Chalo Sansad” march, not a single officer in Delhi has been held accountable. https://t.co/I0YSPsD1IW
India's AI opportunity is not to win the race to the most advanced model, but to show how AI can be made useful to real people solving real problems. India has the ingredients, but @urvashi_aneja writes that a few things stand in the way.
📖Read https://t.co/JEYdUNq1k7
Completely neutral and strictly educational thread 🧵
Non-exhaustive list of all private vendors who provided surveillance tech and other tools of oppression & control to the Delhi Police during the Jantar Mantar youth protests 1/n
Not sure what robotic dogs, extra lethal missiles, and a military control board have to do with celebrating the 80th Indian Independence Day but that's what you'd leave away with if you were watching the PMO livestream on YouTube
real question: where do all the phone numbers you give at almost every juncture in India (upon receiving a courier, when sending a courier, using wifi, ordering something online) go ? Who do they go to? For what purpose?
The World Cup has already given us so much incredible content like this tweet. But, not all posts have been kosher. How are companies dealing with this? I write about it in a new piece: https://t.co/R1QiW2pe4L
What does footy have to do with content moderation? A lot. Between AI-generated Kimmel rants, racist speech targeting French players, and VARgentina conspiracies, T&S teams have their hands full ahead of the World Cup semi finals beginning today: https://t.co/R1QiW2pe4L
What does footy have to do with content moderation? A lot. Between AI-generated Kimmel rants, racist speech targeting French players, and VARgentina conspiracies, T&S teams have their hands full ahead of the World Cup semi finals beginning today: https://t.co/R1QiW2pe4L
New from me and the inimitable @CaitlinVogus for @theintercept. Age verification comes at a great cost to our rights and democracy, including by undermining the anonymity that is crucial for journalists to report on government abuse of power freely https://t.co/hmJMuzx0kc
Mandating age verification "effectively hands Big Tech and the government a skeleton key to the identities of every whistleblower, dissident, and investigative reporter who uses online platforms," FPF's @CaitlinVogus and @AliyaBhatia write in @theintercept.https://t.co/m7aO8U2Hm9
In the new edition of #CODISignals, @CenDemTech's @AliyaBhatia highlights the need for collaboration among industry, academia, civil society and others to advance multilingual AI.
Read more: https://t.co/A59oMFtDEg
"Once identity checks to access websites become normalized, any site can request that personal data and sell it, with little to no ability for users to meaningfully opt out or rectify the situation."
@AliyaBhatia on the downward spiral of age verification: https://t.co/o9BdHa2P0I
AI isn't flattening my speech. Trying to communicate clearly in a world where AI systems are mediating and transforming people's expectations of what is clear written text is changing and homogenizing how I structure my writing, which is actually a more insidious outcome
As AI chatbots begin playing the role of Google Translate, AI companies can and should do more to incorporate linguistically diverse datasets and expertise into model development. Here's what we recommend (as part of our UN Global Dialogue submission) https://t.co/RsU4E3f5nw
Happy birthday to #GoogleTranslate which turns 20 today! Google Translate has helped me in countless situations: from translating my nani'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
At the same time, researchers have pointed out that Google Translate has its limits, particularly when dealing with slang, technical terms and struggling with translations between non-English languages. We cite some of that research here: https://t.co/6yuePGIx59