Shakeel

Shakeel

@shakeelhashim.com

Writing about AI policy at Transformer. Supporting AI journalism with grants at Tarbell. Prev: AI safety and EA comms; journalist at The Economist, Protocol, Finimize. London. https://transformernews.ai

3.6K followers2.7K following250 posts

Shakeel reposted

SCOOP: AI safety group Public First spent in the Manny Rutinel race. It just wasn’t disclosed to voters ahead of election day. Transformer has learned $2m was transferred from Public First Action to Latino Victory Fund, which says it exists to mobilize Latino voters and elect Latinos to office.

transformernews.aiAn AI safety group hid its election spending through a Latino-focused PACPublic First Action routed $2m to support Colorado House candidate Manny Rutinel through Latino Victory Fund — and didn’t publicly announce it ahead of Rutinel’s victory

Shakeel reposted

Happily, you don’t need to do much to trace any of this as we’re extremely transparent about it! www.transformernews.ai/p/standards-...

transformernews.aiEditorial standards and ethicsThis document outlines Transformer’s aims, funding, ethics policies, and editorial standards.
Timnit Gebru@timnitgebru.blacksky.app

Instead of "the left isn't paying enough attention to these machine gods" they're claiming to build, perhaps we should write an article about how effective altruists brand themselves "left", ask who is writing these articles, and trace their cults and sources of funding.

Shakeel reposted

So much interesting in this (and good work on @caseynewton.bsky.social documenting it). One thing to note though, Google's watermarking identified the image of the badge as created by its AI tool. Other image generators haven't bothered implementing anything similar, why? (Looking at you, Sora.)

Olivia Solon@oliviasolon.bsky.social

Great work from @caseynewton.bsky.social and an illuminating case study in how journalists authenticate allegations/sources in an age of generative AI www.platformer.news/fake-uber-ea...

Shakeel reposted

Rob, is it true you created an early language model trained on scraped data? "Mark V Shaney"? Is it also true that it pretended to be a human, sending unsolicited messages on Usenet forums for lonely singles? That's just what I'm seeing here on Wikipedia...

rob pike@robpike.io

Fuck you people. Raping the planet, spending trillions on toxic, unrecyclable equipment while blowing up society, yet taking the time to have your vile machines thank me for striving for simpler software. Just fuck you. Fuck you all. I can't remember the last time I was this angry.

Shakeel reposted

I got the AI writing homies together to launch SAIL (readsail on socials): A team bundle to get access to the best content in AI all in one subscription. Otherwise, nothing in my mission is changing. Get the best content written to inform the public and shape the trajectory of AI. readsail.com

Shakeel reposted

It certainly seems that he'll be able to shape AI policy more than his predecessor, Feryal Clark, did. But he's got an uphill battle ahead, navigating copyright, turf wars with Ed Miliband, and the general complexity of the topic.

Shakeel reposted

Data centers use 2% of the world’s electricity. There’s a *huge* push to build more for AI. But even the most optimistic growth paths for AI predict it’ll drive less growth in energy demand than electric cars or aircon. It’s significant, but it’s not the biggest thing going on.

transformernews.aiWe’re getting the argument about AI's environmental impact all wrongIndividual ChatGPT queries are a rounding error — we need to think about the future
View full thread

Our aim is to be the definitive source for coverage of transformative AI — one dedicated to explaining what’s happening in AI and why it matters. We think this is a really important mission. The world needs more journalism that takes AI seriously. We hope to provide it.

View full thread

We're also introducing an external opinion section, which launches tomorrow with some big names. We hope it'll be a forum for incisive debate about transformative AI from a diverse range of expert perspectives.

View full thread

Shakeel reposted

Why assume Meta’s hiring freeze is for financial reasons, rather than logistical ones (which seems much more likely)? It makes a lot of sense to pause hiring shortly after a huge hiring surge and restructuring — you need time to get the new structure working before bringing more people in!

Rob Cyran@robcyran.bsky.social

Meta was on track to spend $50bn on R&D this year (compare to about $60bn on capex). Quite a turnaround to go from $300m comp packages to hiring freeze in a few weeks.

Hugging Face still has a ton of porn hosted on its website which can be accessed without logging in or verifying your age. Not entirely sure how that's compliant with the UK Online Safety Act ... which even 4Chan's taken action to comply with today!

Shakeel reposted

It's a bit dispiriting to see a company that talks so much about being "ethical" adopt the same old strategy of relying on journalists to find bad stuff on their platforms. I saw this for years covering Meta, and Hugging Face is following exactly the same playbook. bsky.app/profile/shak...

Shakeel@shakeelhashim.com

We sent the list of models we found to Hugging Face. The company removed some, but not all, of the models we shared with them. They left up a video model that references a film released when the actress was 15, for instance.

Here's a story about Hugging Face, deepfake pornography, and irresponsible content moderation. Last week, 404 Media found that Hugging Face hosted 1000s of nonconsensual AI models of real people, many clearly intended to be used to create sexual content.

View full thread

Here's a story about Hugging Face, deepfake pornography, and irresponsible content moderation. Last week, 404 Media found that Hugging Face hosted 1000s of nonconsensual AI models of real people, many clearly intended to be used to create sexual content.

View full thread

Transformer found that many of these models are uploaded by a few prolific users. While Hugging Face removed *some* of those users' uploads, it didn't ban the users' accounts — and today we've found that they're hosting lots of *other* models intended to generate deepfake porn.

We will send these additional URLs to Hugging Face, and perhaps the company will remove them. But a company worth $4.5bn shouldn’t be relying on journalists to keep its platform free of tools that clearly violate its own policies.

Here's a story about Hugging Face, deepfake pornography, and irresponsible content moderation. Last week, 404 Media found that Hugging Face hosted 1000s of nonconsensual AI models of real people, many clearly intended to be used to create sexual content.

And in a Transformer investigation yesterday, we found that many of those models were still online, including at least one instance where a model was explicitly aimed at creating sexual content featuring a teenage-looking likeness of an actress.

We sent the list of models we found to Hugging Face. The company removed some, but not all, of the models we shared with them. They left up a video model that references a film released when the actress was 15, for instance.

Shakeel reposted

It is pretty shameful that Hugging Face is still hosting these models for making deepfake porn of celebrities (and others), despite @404media.co's reporting a week ago.

Shakeel@shakeelhashim.com

Hugging Face portrays itself as the "ethical" AI company. So why is it hosting tools to make deepfake porn of teenage celebrities? Following on from @404media.co's report last week, Transformer found that the company hosts some *really* shocking stuff.

Hugging Face portrays itself as the "ethical" AI company. So why is it hosting tools to make deepfake porn of teenage celebrities? Following on from @404media.co's report last week, Transformer found that the company hosts some *really* shocking stuff.

View full thread

Hugging Face portrays itself as the "ethical" AI company. So why is it hosting tools to make deepfake porn of teenage celebrities? Following on from @404media.co's report last week, Transformer found that the company hosts some *really* shocking stuff.

View full thread

Elsewhere on Hugging Face are dozens of other models explicitly designed to generate pornographic images of female celebrities. These have explicit filenames, and are stored in a folder that also contains models designed to generate photos of them in different sexual positions.

This was ridiculously easy for us to find. Yet Hugging Face haven't taken it down. When asked for comment, Hugging Face head of comms Brigitte Tousignant just told us to report the content.

Hugging Face portrays itself as the "ethical" AI company. So why is it hosting tools to make deepfake porn of teenage celebrities? Following on from @404media.co's report last week, Transformer found that the company hosts some *really* shocking stuff.