\n<\/p>\n
OpenAI is rolling out its latest advanced LLM, Sol, for wide public access. Sol is considered to be at least on par with Anthropic\u2019s Fable, a model whose capabilities (or ownership) stressed out the White House enough that it was briefly banned from public access.<\/p>\n
So how did these models get the OK for release? Short answer: Nobody\u2019s quite sure.<\/p>\n
\u201cFrankly, I don\u2019t have visibility into those exact processes, so yes, I don\u2019t feel like I have enough information to say whether they\u2019re adequate or not,\u201d Mina Narayanan, a senior research analyst at Georgetown\u2019s Center for Security and Emerging Technology, told TechCrunch. \u201cAnthropic did say that they were in conversations with the government, and that they developed a classifier to detect jailbreak attempts, and they\u2019ve implemented defense-in-depth strategies to prevent future jailbreaks, but exactly what that dialog looked like between the government and Anthropic and OpenAI is unclear.\u201d<\/p>\n
Dean W. Ball, a former Trump policy advisor who now works for OpenAI, wrote<\/a> that \u201cnobody knows what the requirements are to get licensed\u201d in his newsletter last month.<\/p>\n Andy Konwinski, a computer scientist who co-founded Databricks, Perplexity, and the Laude Institute, said he\u2019s never spoken to anyone who understands the process, even employees at frontier labs. \u201cIt\u2019s existentially a problem,\u201d he tells TechCrunch. \u201cSafety or not, it\u2019s about who has the power to make decisions \u2014 who gatekeeps and decides on permissions?\u201d<\/p>\n Eighteen months into the Trump administration, there is still little clarity about how to move forward, despite \u2014 or, some critics allege, because \u2014 of the industry figures setting policy. Last month, after weeks of infighting<\/a>, an executive order was published laying out a roadmap for evaluating frontier models, but the specifics have yet to be filled in, other than what won\u2019t exist. \u201cThere will not be an FDA for AI,\u201d Sriram Krishnan, a former Andreessen Horowitz partner who served as a senior advisor for AI in the White House until last month, told<\/a> Financial Times.<\/p>\n Notably, there\u2019s still no agreement on what kinds of models require government scrutiny, or what agency or agencies should perform those evaluations. For now, the Department of Commerce\u2019s Center for AI Standards and Innovation seems to be taking the lead, but the executive order instructs six cabinet agencies to determine a final process by early August. What has emerged in the meantime is, at best, ad hoc.<\/p>\n OpenAI CEO Sam Altman said<\/a> on CNBC that the process involved conversations with the officials like Secretary of Commerce Howard Lutnick, Secretary of the Treasury Scott Bessent, and U.S. national cyber director Sean Cairncross, but it\u2019s not clear who the experts that tested the models were or how they did that. OpenAI declined to share details on the government\u2019s process with TechCrunch, but pointed to the results of several external evaluations by organizations like U.K. AISI, SecureBio, and Irregular in the latest model\u2019s safety card<\/a>.<\/p>\n As with Anthropic\u2019s Fable roll-out, OpenAI previewed the model for the government and select users ahead of wider release, but we don\u2019t know who all of those users were or how they were chosen. In a late June blog post<\/a>, the company said \u201cwe don\u2019t believe this kind of government access process should become the long-term default,\u201d saying it would work with the government to develop a different path forward.<\/p>\n The backdrop to those conversations, however, includes Altman reportedly offering<\/a> as much as 5% to OpenAI\u2019s equity for the administration\u2019s so-called \u201cTrump Accounts,\u201d and OpenAI president Greg Brockman\u2019s role as the largest publicly known donor<\/a> to Trump\u2019s mid-term political operation. It\u2019s hard for outside observers to separate those activities from the government\u2019s apparently lighter-touch approach to regulating Sol.<\/p>\n Anthropic\u2019s Fable, on the other hand, was briefly pulled from wider access when the U.S. government forbade its use by foreign nationals, partly because of real concerns about users jail-breaking the model to access hacking capabilities and partly due to personality clashes between Anthropic and the Trump administration. The threat of an export ban may have also led OpenAI to be more cooperative with the government\u2019s (unknown) requests.<\/p>\n From an industry perspective, a hands-off approach to regulation might be nice, but one that depends on personal connections to administration officials creates uncertainty and bad incentives. <\/p>\n Konwinski told TechCrunch that he worries true experts in this technology \u2014 \u201csafety researchers, alignment researchers, interpretability researchers, but also data people, and people from all over the stack\u201d \u2014 aren\u2019t playing enough of a role in the model release process.<\/p>\n Konwinski argues<\/a> that an \u201copen commons\u201d is the best way to actually balance safety and innovation. He points to models like the FDA, the NIH, or the national labs, which convene researchers, government officials, and private companies to reach a consensus on safety issues.<\/p>\n Some of that comes down to the incentives of capitalism that have motivated AI researchers for more than a decade, and played out in the court room during Elon Musk\u2019s lawsuit challenging OpenAI\u2019s corporate structure. Ball points out that the nature of the AI business requires companies to recoup much of their training costs shortly after their models are released and are further ahead of the competition.<\/p>\n \u201cEven if their intentions are good, there\u2019s very clear legal obligations and fiduciary responsibility that are built right into the operating procedures,\u201d Konwinski points out.<\/p>\n Ball, in his post<\/a>, argued that the way forward will depend on third-party auditing organizations, licensed by the government, that will evaluate frontier labs\u2019 approach to safety. Konwinski, too, is bullish about new institutional formats like focused research organizations that could help more disinterested experts from academia and the nonprofit world access and evaluate frontier models.<\/p>\n For now, the secrecy around the development of AI isn\u2019t going away, but it also will seed political challenges for an industry that Americans increasingly view with skepticism<\/a>. \u201cThere\u2019s not a sense that responsible people are driving forward these changes,\u201d University of Wisconsin-Madison computer science professor Remzi Arpaci-Dusseau said last week at the Open Frontier<\/a> conference.<\/p>\n At the same event, David Siegel, the computer scientist who founded Two Sigma, one of the most successful quantitative hedge funds, asked attendees to \u201cimagine a situation, which I think would be very bad, [where] a small number of firms control the technology; the government, in their secretive laboratories, is evaluating whether or not the technology is suitable for use; and the general public and scientific community doesn\u2019t really have any access to any of that stuff.\u201d<\/p>\n It seems like we don\u2019t need to imagine it.<\/p>\n<\/div>\n When you purchase through links in our articles, we may earn a small commission<\/a>. This doesn\u2019t affect our editorial independence.<\/em><\/p>\n