Occupy Wall Street Co-Founder Built an AI App to Help Activists Seize the Means of Computation

Occupy Wall Street co-founder Micah White has launched Outcry, a new AI chatbot designed as a private, on-device mentor for activists and organizers. The app aims to provide left-wing movements with specialized, high-quality AI tools, countering what White sees as the increasingly conservative nature of the tech industry. Outcry represents a novel effort to repurpose large language models for progressive political organizing.

In an era where Silicon Valley’s conservatism is both expressed openly and becoming more intense by the day, it’s strange to think that tech was once seen as a hive of liberalism. The right-wing nature of today’s tech industry means that its products tend to also be seen as serving right-wing interests, either in their actual operation like X’s openly and unrepentantly right-wing chatbot Grok https://gizmodo.com/grok-is-a-flop-but-it-may-not-matter-to-elon-musk-2000757570 or by the simple fact that their existence serves to enrich a small group of very powerful, very conservative people. But does it have to be this way? Can LLMs and AI agents find a place in the toolkit of progressive activist groups? The conviction that they can is the idea behind a new app called Outcry https://apps.apple.com/us/app/outcry-activist-ai-mentor/id6762086768 , which provides a chatbot designed specifically as a “private, on-device AI mentor for activists, organizers and movement builders.” There’s also a web version https://www.outcryai.com/ , although it obviously lacks the privacy benefits of being entirely offline. It’s the brainchild of Occupy Wall Street co-creator Micah White, who recently wrote a blog post https://micahbornfree.substack.com/p/i-crammed-an-activist-ai-into-an about the thinking behind the project. This is actually a fascinating idea. It’s certainly not the only way in which activist groups are leveraging Big Tech’s own products against it. We reported last month https://gizmodo.com/some-locals-are-using-ai-to-protest-against-data-centers-2000744027 on the ways in which locals are using LLMs to assist in their campaigns against the construction of data centers. But it’s the first example that I’m aware of in which an LLM has been designed specifically for the use of left-wing activists. In addition, it’s also an example of an increasingly common use for LLMs—as specialist advisers. Ultimately, LLMs are databases https://www.youtube.com/watch?v=8Ppw8254nLI : they’re a fancy, natural-language way of querying large datasets for the information you want. Generalist chatbots like ChatGPT are trained on pretty much anything and everything. In theory, this means you can ask them about quantum chromodynamics just as easily as you can ask them to provide advice on how to roast a chicken. In practice, it creates a huge problem with distinguishing reliable information from nonsense, because otherwise there’s nothing to distinguish a peer-reviewed study on the efficacy of mRNA vaccines from a bunch of cookers in a Facebook group wringing their hands about how the Covid shot has given them Morgellon’s, or something. However, if you restrict the training data to a given subject and to sources chosen specifically for their reliability and relevance, you can create something like Outcry: a specialized chatbot that you know has been trained on high-quality data. This isn’t to say that the app can’t make mistakes—the nature of LLMs is that they’ll try to collate various pieces of data into a coherent, natural-language result, and given that they’re ultimately relying on their internal weighting parameters and pattern recognition algorithms, sometimes the result they spit out will be inaccurate. This is separate to the problem of LLMs spontaneously generating fake information, commonly referred to—inexplicably and irritatingly—as “hallucinating.” This certainly isn’t the first application of this idea. It was reported https://www.nbcnewyork.com/news/national-international/most-us-doctors-use-this-ai-tool-few-patients-know/6501320/ earlier this month that something like two thirds of doctors ask a specialized medical LLM called OpenEvidence for advice in making diagnoses. This isn’t necessarily a bad thing; doctors certainly Googled things before the advent of LLMs. It really just depends where they’re asking, and whether they’re checking the responses they get to make sure they’re factually correct. Having said that, if you see your doctor asking ChatGPT about your symptoms, you should probably find a new doctor. Outcry’s other distinguishing feature is that its dataset is entirely offline—it’s included with the download. According to the readme, the entire dataset is downloaded to your device at first launch, and stored in your library’s Application Support directory. I thought it might be interesting to have a look at the data, but despite all manner of grep-ing and rummaging through both my user library and the computer’s main library, I couldn’t find the file in question. It doesn’t seem to be in the download package, either, despite the weighty 3GB download. Asking the chatbot where its data was stored also got me nowhere. It’s clearly somewhere, though, because the app itself is clear about its offline nature. So with all that said, how does Outcry fare in its role as an organizing mentor? I’d say that its information is pretty high-level and general, not least because its offline nature prevents it from accessing specific details not contained in its database. Take, for example, its answer about local anti-ICE groups in NYC: With that said, with a bit of prompting it’ll suggest concrete actions you can take, and sometimes do so straight off the bat: According to White, the Outcry app is “imperfect,” and he’s asking activists to test it out and let him know what works so they can improve it. So is Outcry any good? On the whole, I’d say “yes.” This app has the potential to be a really valuable resource, especially for people who are just beginning to become involved with activism and genuinely don’t know where to begin—and getting over that first step can be hard.