The Distributed AI Research Institute (DAIR)

i met with some of Africa's brightest minds in artificial intelligence at this year's annual AI conference Deep Learning Indaba ('Indaba' is a Zulu word for gathering) held in Dakar, Senegal. and I wrote a feature about it.? the?Deep Learning Indaba is a bold AI movement, empowering African researchers who want to make AI that actually works for Africa. but it's not that easy. For MIT Technology Review, I chronicle the significance of the Indaba, the challenges African researchers face and the strides they have made in building AI systems that answer the needs of Africans. a major issue that I examined in this story is the tension around efforts to develop the regulatory framework that'll govern the approach to AI across all of Africa. researchers are in conflict over a continent-wide strategy -- and what policies will most benefit Africa, not the wealthy, Western governments and corporations that have often funded technological innovation in the continent. but the issue transcends creating a single strategic document. all of this speaks to a broader tension over foreign influence in the African?AI scene, including angst about foreign funding for AI development, and skepticism of exploitative, profit-driven big tech companies.? taken together with inadequate local investment, poor infrastructure, and linguistic?challenges, Africa needs to overcome all these barriers to pave its own pathway in the global AI race. as the respected AI maven, Ethiopian computer scientist, Timnit Gebru, told me: "Africans [need] to do something different and not replicate the same issues we're fighting against."? read full story:???https://lnkd.in/ec44tVms?

What does data worker-centric regulation look like? Tomorrow at the European Parliament MEP Le?la Chaibi, Prof. Antonio A. Casilli, and I will present our policy report on data work in Europe. You can read it at https://lnkd.in/dymSaFk8. We'll be joined by data workers from Italy, Spain, Venezuela, Syria, and the US to discuss the labor behind AI and what regulation data workers need: OSKARINA FUENTES ANAYA, Yasser Alrayes, Nacho Barros, Donatella Delpiano, Krystal Kauffman, Dylan Baker. ?? Watch the session live at 2:30pm CET: https://lnkd.in/d-Pz-gXm

A Guide for Creating and Documenting Language Datasets with?Data Statements Schema Version 3 (Angelina McMillan-Major and Emily M. Bender, 2024) is now available! You can find the Guide along with templates for data statements and other resources here: https://lnkd.in/gdD8fiDi Data Statements schema version 3 builds on McMillan-Major's 2023 dissertation Language Dataset Documentation Design: Learning from Deaf and Indigenous Communities. The updated guide serves as both a resource for writing data statements and as an aid for planning data collection.

A quick newsletter post on the dehumanization behind Satya Nadella's remarks about copyright law: https://lnkd.in/g-Jd5rqV

A quick newsletter post on the dehumanization behind Satya Nadella's remarks about copyright law: https://lnkd.in/g-Jd5rqV

?During the negotiations for the AI Act, how many times were the data workers training AI invited at the European Parliament? Zero. With MEP Le?la Chaibi & Dr. Milagros Miceli on Nov 21, we'll be welcoming in Brussels workers and organizers from ????????????????????. Join us live or on YouTube.

What is the status of funding for work addressing Online Gender-Based Violence — what is being funded, what are the gaps, what are our movement aspirations for being funded in flexible, long term and meaningful ways that can help us address OGBV/TFGBV in creative and needed ways? We invite you to contribute to this mapping effort by completing this survey https://lnkd.in/gRQ4BHCk by December 10, 2024. Please share widely to your networks. Survey is in English, Arabic, Chinese, French, Hindi, Portuguese, Russian, Spanish and Tamil.?

Tomorrow!

On their show Mystery AI Hype Theater 3000, Alex Hanna, Ph.D. and Emily M. Bender pick through artifacts washed up in the flood of AI hype and examine them to separate fact from fiction. I spoke to them about the effort on the latest Tech Policy Press podcast:

Mystery AI Hype Theater Episode 44: OpenAI's Ridiculous 'Reasoning' https://lnkd.in/ge-Zcv_R In which Alex Hanna, Ph.D. and I bring a little actual science to the party to cut through the hype. Thx to Christie Taylor for production! Also available as video on PeerTube: https://lnkd.in/gKx6RcRx Alttext for video: Audiogram with Mystery AI Hype Theater 3000 logo, title: "Episode 44: OpenAI’s Ridiculous ‘Reasoning’” and subtitle “How o1’s ‘complex chains of thought’ are assessed with meaningless metrics -- and erase what thinking really is” EMILY M. BENDER: The funniest thing--so this is where they're testing it against the International Olympiad in Informatics problems, I guess. So they're doing some kind of supervised learning effectively over this specific task. And then they say, "This model competed in the 2024 IOI under the same conditions as the human contestants." But, "It had 10 hours to solve six challenging algorithmic problems and was allowed 50 submissions per problem." But the thing that cracked me up about this is, "With a relaxed submission constraint, we found that model performance improved significantly. When allowed 10,000 submissions per problem, (laughter) the Model achieved a score above the gold medal threshold even without any test time selection strategy." ALEX HANNA: I too would like to have 10,000 submissions per every exam I am taking. EMILY M. BENDER: I would never want to be the one grading that exam. So SJayLett says in quotes, "'Under the same conditions.' It had an itchy nose?" ALEX HANNA: Yeah, this whole thing is a bit bizarre. And it's sort of, the analogy of chain of thought to sort of steps of human reasoning is, you know, a further kind of devaluing of what it means for humans to reason and how humans do it. EMILY M. BENDER: SJayLett adds, "I suspect it emphasizes how the people behind this don't really believe the mind is embodied in any meaningful way," which I think is very true.