A new report from Politico dives into the inner workings of a billionaire-backed network of AI fellows in key congressional offices that are helping to shape AI policy. Now, the key player in this is an organization called Open Philanthropy. They're self-described philanthropic funder whose mission is to help, [00:00:19] according to them, help others as much as we can with the resources available to us. So little bit from that excellent reporting from Politico. The fellows funded by Open Philanthropy, which is financed primarily by billionaire Facebook co-founder and [00:00:34] Asana CEO Dustin Moskowitz and his wife Cari Tuna, are already involved in negotiations that will shape Capitol Hill's accelerating plans to regulate AI. Now, they've been acting through the Horizon Institute for Public Service, which is a nonprofit that Open Philanthropy has [00:00:53] effectively created in 2022. Now, the group is funding the salaries of tech fellows in some key senate offices. So a little bit more on that. Current and former Horizon AI fellows with salaries funded by Open Philanthropy are now working at the Department of Defense, [00:01:09] the Department of Homeland Security, and the State Department, as well as in the House Science Committee and Senate Commerce Committee, two crucial bodies in the development of AI rules. They also populate key think tanks shaping AI policy, including the RAND Corporation and Georgetown University's Center for [00:01:26] Security and Emerging Technology, according to the Horizon website. In 2022, Open Philanthropy set aside nearly $3 million to pay for what ultimately became the initial cohort of Horizon fellows. So now, Open Philanthropy has attempted to convince people that they're [00:01:45] entirely separate from Horizon, but it's very obvious that the organization is just using Horizon as an attempt to mask the fellows' ties to the company. Now, because Horizon is considered a nonprofit, the money that Open Philanthropy has spent on Horizon fellows since its initial grant [00:02:02] is something that we can't know. Now, this opens the door to multiple conflicts of interest. A little bit more on that. Tim Stretton, director of the congressional oversight initiative at the Project On Government Oversight, said congressional fellows should not be allowed to work on issues where their [00:02:19] funding organization has specific policy interests at play. He added that fellows should not draft legislation or educate lawmakers on topics where their backers conceivably stand to gain a dynamic apparently at play in the case of Horizon's fellowship program, given Open Philanthropy's ties to OpenAI and Anthropic. [00:02:37] Now, when asked about the ethical and conflict of interest issue, Horizon co-founder and executive director, Remco Zwetsloot said the fellowship program is not for the pursuit of particular policy goals, [00:02:53] does not screen applicants for belief in long-term AI risks, and includes fellows with a diverse set of views on AI's existential dangers. And we're talking about AI, so this might be a little bit complicated. So I'll just sort of make this analogy that the Politico article made when [00:03:11] speaking this, what I think quite succinctly, they likened it to. If APAC had a nonprofit wing of its organization that was funding Middle Eastern foreign policy fellows in congressional offices, [00:03:26] that would be clearly a conflict of interest. But that is precisely what is happening here. So I'll open it to you guys. Now, we got a little bit more, but for your thoughts on this. >> Speaker 2: All right, thank you for running the story because it gives me [00:03:42] an opportunity to talk about what's actually going on with AI. So these are comments I've made multiple times on the Young Turks and other networks. The AI systems that generally are being spoken about today are not traditional AI systems. They're using our own personal data collected by large server farms to mine [00:04:00] and process that data because the storage and processing of data has become exponentially cheaper over time, which is an amazing technical development. They're using all of our personal data to basically create systems that can emulate human behavior, that can mimic human behavior. [00:04:18] So that's just what that kind of said. Anytime you see words, especially in relation to tech, like open, or philanthropy, or innovation, or disruption, that's a sign usually that they're trying to brand you out and bait and switch you, right? [00:04:36] So the term open is like the gospel in tech worlds, which I come out of in some ways. OpenAI is not open because they don't disclose what data is being gathered about us or what they're doing with that data, how long they're retaining it for, and how their business models associatedly. [00:04:56] Similarly here, okay, so what that is doing is actually masking the actual tasks that need to be done to actually regulate AI so it serves people, laborers, our planet, and so on. And the other thing that they're doing and this story alludes to it is, it's a common [00:05:16] thing we've seen, especially with the emergence of GPT and OpenAI and so on. We're the only ones who can regulate these things because we're the only ones who know what's actually going on. Never mind that we call ourselves open. We're the only ones who actually know what's going on. We, being tech, CEOs, investors, those people. [00:05:35] And that's because intentionally, these systems are built in opaque ways based on the surveillance and capture of personal data. They have intentionally built systems that order the world for us in ways that 99.9% of us have no visibility or accountability over. [00:05:53] So, that's why you hear about the terminator scenario, which is there's an existential risk. That's why you hear about, we're the ones best left to regulate all of this ourselves. That all masks the realities of what these AI systems are doing, [00:06:08] which is threatening workers at a time, where thankfully, we have more unionization energy than recent history from what I can remember. They mask the racist and incarcerating aspects of these technologies. They're being used in prisons, police departments, and so on. [00:06:26] They're misidentifying people who are as criminals, who are not criminals, because they're merely black. These are all examples of what we could call algorithmic violence or AI-level violence. So, every time one of these stories like this drops, and I'm so glad the Young Turks covers stories like this, [00:06:43] we have to actually look at beneath the veneer and all the branding, like fancy branding language that tech companies always use, what is actually at stake. It's a threat to the working class and a threat to our own personal data. Instead, you can imagine how machine learning systems, including large language [00:07:00] model systems like ChatGPT, could actually serve everybody. That would take actual regulation because that would involve people who actually care about the working class, about citizens, about democracy, and so on. The main existential risk is the possibility of building an AI system that [00:07:16] can do mass-scale disinformation. And that is a realistic scenario, but that's kind of already happening, by the way. So I think the main thing here to recognize is they are shielding themselves from actual regulation. So what if the National Labor Relations Board was the actual regulator of these [00:07:32] systems? What if the FDC was the actual agent that could actually change these systems? I think there's a lot of possibility to transform the relationship between systems of all kinds and our democracy and our economy to support everybody. [00:07:48] But that is absolutely not what's happening. So just don't believe the hype. Don't believe the hype when it comes to what tech companies always push out upon us, because they're not tech companies. They're the wealthiest and most powerful companies in many cases in the history of the world. Many of them are military contractors, too. [00:08:05] And that's very worth noting. >> Speaker 3: Yeah, that last part especially is deeply, deeply concerning. Ramesh knows this pretty much as well as anybody does. So I'll just comment on, look, even if it's philanthropy or it's a nonprofit, first of all, a lot of things are disguised as nonprofits that are nefarious. [00:08:23] But even if they meant well, you cannot allow outside forces to embed inside the government. As Rayyvana mentioned, APAC's theoretically a nonprofit, right? But they are for the benefit of a different country. [00:08:40] But you could pick anything. Saudi Arabia has tons of nonprofits. Are we going to let them embed inside the United States government? That's crazy, right? And so I'm worried about the wholesale takeover of our government. And there's another thing that is, in a sense, [00:08:58] AI that has already taken over the government. It's called corporations. So we created corporations like we created AI, and we wrote the code for corporations. And the code we wrote was disastrously wrong. We wrote only one line of code, maximize profit. [00:09:16] And when you do that, and you let that AI machine run wild, it maximizes profit at the cost of everything in its path, including us, the humans that created those corporations. And what did they do? [00:09:31] They eventually figured out how to buy the Supreme Court. And this is exactly the core of our problem that I explained in the book that I just wrote, Justice Is Coming. And after they'd taken over the Supreme Court, the Supreme Court allowed for bribery, which then allowed them to take over all of the rest of government. [00:09:50] Now, the embeds they have in our government are the senators and representatives themselves. So now here comes new embeds within those embeds. But the only people not represented in all of this is us, the actual human beings that this democracy is supposed to serve. [00:10:07] There's something deeply wrong in Washington. We have to take our government back from all of these things that we created that have now run amok and are now controlling us rather than the other way around, which is how it was supposed to be.