OpenAI's advisory board calls for continued and strengthened nonprofit oversight
That is the message from an advisory board convened by OpenAI to give it recommendations about its nonprofit structure — delivered in a report released Thursday, along with a sweeping vision for democratizing AI and reforming philanthropy.
'We think it's too important to entrust to any one sector, the private sector or even the government sector,' said Daniel Zingale, the convener of OpenAI's nonprofit commission and a former adviser to three California governors. 'The nonprofit model allows for what we call a common sector,' that facilitates democratic participation.
The recommendations are not binding on OpenAI, but the advisory commission, which includes the labor organizer Dolores Huerta, offers a framework that may be used to judge OpenAI in the future, whether or not they adopt it.
In the commission's view, communities that are already feeling the impacts of AI technologies should have input on how they are developed, including how data about them is used. But there are currently few avenues for people to influence tech companies who control much of the development of AI.
OpenAI, the maker of ChatGPT, started in 2015 as a nonprofit research laboratory and has since incorporated a for-profit company with a valuation that has grown to $300 billion. The company has tried to change its structure since the nonprofit board ousted its CEO Sam Altman in Nov. 2023. He was reinstated days later and continues to lead OpenAI.
It has run into hurdles escaping its nonprofit roots, including scrutiny from the attorney generals in California and Delaware, who have oversight of nonprofits, and a lawsuit by Elon Musk, an early donor to and founder of OpenAI.
Most recently, OpenAI has said it will turn its for-profit company into a public benefit corporation, which must balance the interests of shareholders and its mission. Its nonprofit will hold shares in that new corporation, but OpenAI has not said how much.
Zingale said Huerta told the commission their challenge was to help make sure AI is a blessing and not a curse. To grapple with those stakes, they envision a nonprofit with an expansive mandate to help everyone participate in the development and trajectory of AI.
'The measure of this nonprofit will be in what it builds, who it includes, and how faithfully it endures to mission and impact," they wrote.
The commission toured California communities and solicited feedback online. They heard that many were inspired by OpenAI's mission to create artificial intelligence to benefit humanity and ensure those benefits are felt widely and evenly.
But, Zingale said many people feel they are in the dark about how it's happening.
'They know this is profoundly important what's happening in this 'Age of Intelligence,' but they want to understand better what it is, how it's developed, where are the important choices being made and who's making them?' he said.
Zingale said the commission chose early on not to interact with Altman in any way in order to maintain their independence, though they quote him in their report. However, they did speak with the company's senior engineers, who they said, 'entered our space with humility, seriousness, and a genuine desire to understand how their work might translate into democratic legitimacy.'
The commission proposed OpenAI immediately provide significant resources to the nonprofit for use in the public interest. For context, the nonprofit reported $23 million in assets in 2023, the most recent year that its tax filing is available.
The commission recommend focusing on closing gaps in economic opportunity, investing in AI literacy and creating an organization that is accessible to and governed by everyday people.
'For OpenAI's nonprofit to fulfill its mandate, it should commit to more than just doing good - it should commit to being known, seen, and shaped by the people it claims to serve,' they wrote.
The commission suggested opening a rapid response fund to help reduce economic strains now. Zingale said they specifically recommended funding theater, art and health.
'We're trying to make the point that they need to dedicate some of their resources to human to human activities,' he said.
The commission also recommend setting up a requirement that a human lead the nonprofit, which Zingale said is a serious recommendation and 'a sign of the times."
___
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Tom's Guide
21 minutes ago
- Tom's Guide
Sam Altman gives stern warning on AI, fraud and passwords — 'That is a crazy thing to still be doing'
Sam Altman, CEO of OpenAI, has quickly become one of the biggest names in AI. Like some of his competitors, Altman's position in AI hasn't stopped him speaking out on concerns in the industry. Most recently, that's around the AI fraud crisis. With AI and deepfakes, it is easier than ever for people to impersonate others and commit fraud using these advanced tools. 'A thing that terrifies me is apparently there are still some financial institutions that will accept a voice print as authentication for you to move a lot of money or do something else — you say a challenge phrase, and they just do it,' Altman said at the Federal Reserve conference in Washington. That is a crazy thing to still be doing… AI has fully defeated most of the ways that people authenticate currently, other than passwords 'That is a crazy thing to still be doing… AI has fully defeated most of the ways that people authenticate currently, other than passwords.' These comments from Altman come as AI shows its ability to replicate voices more efficiently than ever. While safeguards are often put in place by some of the bigger companies, it is still a simple process to replicate anyone's voice with a few short recordings. The technology is also growing rapidly in the video world. There are concerns that, in the near future, we'll be able to easily produce a video and audio combination that can fool anyone into believing they are having a full conversation with a real person. This speech comes as the White House ramps up to release its 'AI action plan'. This would be a policy document that outlines its approach to AI regulation. This is something other countries have looked to do as well, but few have produced very detailed approaches. Get instant access to breaking news, the hottest reviews, great deals and helpful tips. OpenAI is gearing up to be ahead of the AI movement in terms of policy. The company has confirmed that it will soon open its first Washington, DC office with a small workforce. This would be used to host policymakers and provide AI training to those making decisions on the technology. The company also recently announced an agreement with the U.K. government to find ways to use AI in government decisions, as well as helping explain the benefits and concerns of AI. Altman isn't alone in concern from the higher-ups of AI. Anthropic's CEO Dario Amodei has recently warned about the effects of AI on job security, and the Future of Life institute has recently claimed that most AI companies are not planning enough for the future. However, where AI can be a risk, it can also be a benefit. While some AI tools look to create scams, others look to detect it. As the technology gets better, it becomes more useful on both sides of the spectrum.
Axios
an hour ago
- Axios
What news sources AI chat bots read
News stories or content generated by external sources like journalists, influencers, customers, or the general public are the top sources for AI bots like ChatGPT, Claude and Gemini, a new Muck Rack report finds. Why it matters: AI is increasingly being used for search, and how a brand, company or public figure shows up in AI-generated responses could impact their ability to attract customers, investors and talent. By the numbers: Muck Rack input more than 1 million realistic user prompts into ChatGPT, Gemini and Claude and analyzed the citations. 96% of links cited by AI sit squarely within the purview of communications and corporate affairs, per the report. 37% of inquiries cite external blogs or content (those not owned by the company or product targeted in the query) and 27% cite news stories produced by journalistic entities. 9% cite owned content and government or NGO websites, 7% cite aggregators and 6% cite academic research. Social media and marketing content each make up 2% of the results, while press releases make up 1%. Between the lines: A key difference between AI-generated search results and traditional SEO is that paid marketing and sponsored links rarely populate. "Based on the data, we can see that the models are pretty clearly avoiding marketing materials," says Matt Dzugan, senior director of data at Muck Rack. "Essentially, the models are trying to earn the trust of [their] audience and don't want to regurgitate salesy materials." Instead of paid marketing, owned content like thought leadership, fact sheets or corporate blogs "seem to be the sweet spot for for getting your content cited by these models," he added. State of play: Fact-based queries and prompts are more likely to cite news outlets in the responses, the report found. The outlets most cited include Reuters, the Financial Times, Time, Axios, Forbes and the Associated Press. Of note, most of these outlets have publisher partnerships with OpenAI, however many are also cited regularly across other LLMs. Yes, but: New media, like Substack newsletters and podcasts, are showing up too, but in more indirect ways. As of now, this content is populating through social citations. For example, if "the podcast is on YouTube, and then YouTube generates a transcription file, then Google, in particular, will cite it," says Dzugan. "If you're a PR person evaluating which podcasts to go on, it's important to know that if you go on one that also publishes on YouTube, you have a better shot at influencing Google Gemini than you would with an [audio only] podcast," Muck Rack CEO Greg Galant added. LinkedIn, Reddit and Glassdoor — places where user-generated content and reviews can be found — can also influence an LLM's response, he added. Driving the news: To handle this new reality, Muck Rack will launch Generative Pulse, a tool the monitors how brands are represented in generative AI platforms, says Galant. "In this new GEO [generative engine optimization] world, recent content or news stories are what's driving the answers," says Galant. "Journalism and [third-party] articles really affect the outcome, and that impacts the work of the PR function much more than marketing," he added. Zoom out: Citations can change based on how the query is asked and each LLM has its own habits when responding, says Dzugan. For example, the report found that Claude uses media outlets the least and is more likely to pull from academic, federal and technical sources. ChatGPT is the heaviest user of news citations, pulling from mainstream publications like Reuters, AP, FT, Time and Axios. Zoom in: Niche websites, trade publications and web encyclopedias are also permeating AI-generated responses, the report finds. Wikipedia is cited most across all industry-specific inquiries, according to the Muck Rack report. Scientific and medical databases are frequently cited in AI-generated responses relating to health care and education, while government sites are most likely to inform responses related to energy. Inquiries about government will see federal agencies cited in responses from Claude and Gemini, while ChatGPT incorporates news sources as well. The bottom line: These citations are not random, says Dzugan. "If you study it, you can find the patterns and identify [the] niche journalists, publications or even which Wikipedia pages to prioritize within your industry." "It's on the comms professionals to learn the patterns and then take action on them," he added.
Business Insider
an hour ago
- Business Insider
Here's the list of websites gig workers used to fine-tune Anthropic's AI models. Its contractor left it wide open.
An internal spreadsheet obtained by Business Insider shows which websites Surge AI gig workers were told to mine — and which to avoid — while fine-tuning Anthropic's AI to make it sound more "helpful, honest, and harmless." The spreadsheet allows sources like Bloomberg, Harvard University, and the New England Journal of Medicine while blacklisting others like The New York Times and Reddit. Anthropic says it wasn't aware of the spreadsheet and said it was created by a third-party vendor, the data-labeling startup Surge AI, which declined to comment on this point. "This document was created by a third-party vendor without our involvement," an Anthropic spokesperson said. "We were unaware of its existence until today and cannot validate the contents of the specific document since we had no role in its creation." Frontier AI companies mine the internet for content and often work with startups with thousands of human contractors, like Surge, to refine their AI models. In this case, project documents show Surge worked to make Anthropic's AI sound more human, avoid "offensive" statements, and cite documents more accurately. Many of the whitelisted sources copyright or otherwise restrict their content. The Mayo Clinic, Cornell University, and Morningstar, whose main websites were all listed as "sites you can use," told BI they don't have any agreements with Anthropic to use this data for training AI models. Surge left a trove of materials detailing its work for Anthropic, including the spreadsheet, accessible to anyone with the link on Google Drive. Surge locked down the documents shortly after BI reached out for comment. "We take data security seriously, and documents are restricted by project and access level where possible," a Surge spokesperson said. "We are looking closely into the matter to ensure all materials are protected." It's the latest incident in which a data-labeling startup used public Google Docs to pass around sensitive AI training instructions. Surge's competitor, Scale AI, also exposed internal data in this manner, locking the documents down after BI revealed the issue. A Google Cloud spokesperson told BI that its default setting restricts a company's files from sharing outside the organization; changing this setting is a "choice that a customer explicitly makes," the spokesperson said. Surge hit $1 billion in revenue last year and is raising funds at a $15 billion valuation, Reuters reported. Anthropic was most recently valued at $61.5 billion, and its Claude chatbot is widely considered a leading competitor to ChatGPT. What's allowed — and what's not Google Sheet data showed the spreadsheet was created in November 2024, and it's referenced in updates as recent as May 2025 in other documents left public by Surge. The list functions as a "guide" for what online sources Surge's gig workers can and can't use on the Anthropic project. The list includes over 120 permitted websites from a wide range of fields, including academia, healthcare, law, and finance. It includes 10 US universities, including Harvard, Yale, Northwestern, and the University of Chicago. It also lists popular business news sources, such as Bloomberg, PitchBook, Crunchbase, Seeking Alpha, and PR Newswire. Medical information sources, such as the New England Journal of Medicine, and government sources, such as a list of UN treaties and the US National Archives, are also in the whitelist. So are university publishers like Cambridge University Press. Here's the full list of who's allowed, which says that it is "not exhaustive." And here's the list of who is banned — or over 50 "common sources" that are "now disallowed," as the spreadsheet puts it. The blacklist mostly consists of media outlets like The New York Times, The Wall Street Journal, and others. It also includes other types of sources like Reddit, Stanford University, the academic publisher Wiley, and the Harvard Business Review. The spreadsheet doesn't explain why some sources are permitted and others are not. The blacklist could reflect websites that made direct demands to AI companies to stop using their content, said Edward Lee, a law professor at Santa Clara University. That can happen through written requests or through an automated method like Some sources in the blacklist have taken legal stances against AI companies using their content. Reddit, for example, sued Anthropic this year, saying the AI company accessed its site without permission. Anthropic has denied these claims. The New York Times sued OpenAI, and The Wall Street Journal's parent, Dow Jones, sued Perplexity, for similar reasons. "The Times has objected to Anthropic's unlicensed use of Times content for AI purposes and has taken steps to block their access as part of our ongoing IP protection and enforcement efforts," the Times spokesperson Charlie Stadtlander told BI. "As the law and our terms of service make clear, scraping or using the Times's content is prohibited without our prior written permission, such as a licensing agreement." Surge workers used the list for RLHF Surge contractors were told to use the list for a later, but crucial, stage of AI model training in which humans rate an existing chatbot's responses to improve them. That process is called "reinforcement learning from human feedback," or RLHF. The Surge contractors working for Anthropic did tasks like copying and pasting text from the internet, asking the AI to summarize it, and choosing the best summary. In another case, workers were asked to "find at least 5-10 PDFs" from the web and quiz Anthropic's AI about the documents' content to improve its citation skills. That doesn't involve feeding web data directly into the model for it to regurgitate later — the better-known process that's known as pre-training. Courts haven't addressed whether there's a clear distinction between the two processes when it comes to copyright law. There's a good chance both would be viewed as crucial to building a state-of-the-art AI model, Lee, the law professor, said. It is "probably not going to make a material difference in terms of fair use," Lee said.
