Why the AI CEOs Really Want to Be Regulated

Calls to regulate the AI sector have reached a fever pitch over the last week. Fear-mongering about the prospects of a totally unregulated AI industry has been around as long as the industry itself. But the current storm began last Tuesday when an engineer at Anthropic, the firm best known for its chatbot Claude, resigned citing safety concerns. The engineer claimed that many people at top AI firms earnestly believe the technology could wipe out humanity by the end of the 2020s. Those claims got a boost on Saturday when Dario Amodei, the CEO of Anthropic, published an essay calling for an across-the-board slowdown of AI development. Other prominent CEOs like Sam Altman and Elon Musk then said they agreed with Amodei. The media, which is always on the lookout for a new mass extinction scare, jumped on what could appear to anyone who didn't actually read Amodei's essay as confirmation from the top AI companies that those warning of an AI apocalypse are right. Groups and commentators from all over the political spectrum then jumped into the fight. The topic dominated the news cycle, and now the top AI CEOs are reportedly heading to the White House for a meeting about AI risks. There's a lot going on here, but almost none of it truly has anything to do with our safety. There are, to be sure, people out there who have genuine concerns about the risks posed by the accelerating pace of AI development, but it is hard to believe that that's anywhere close to the primary motivation for the sudden push for more regulations from the chief executives at the companies currently dominating the AI sector. I say sudden, but really this isn't new. It's only just gotten a lot more attention. The truth is many of these top AI CEOs have been pushing for greater regulation of their industry for some time now, and their efforts have escalated in recent months. Over the summer, Amodei, Altman, and Demis Hassabis, the CEO of Google's DeepMind, all published written calls for the creation of remarkably similar AI regulatory regimes. A few weeks later, Meta's Mark Zuckerberg released a lengthy letter where he also argued that AI needs to be regulated globally. Now, according to progressives, this should be considered surprising. We're often told that the billionaire CEOs of large companies spend millions of dollars buying politicians and brainwashing voters to prevent the government from regulating their industries. Their aim is to ensure that the government leaves them entirely alone and allows them to do whatever they want to us, transferring all power in society from the people to the market that these big corporations control. And we libertarians or free market advocates are supposedly either usefully stupid unwitting enablers or nefarious collaborators of the billionaire class. So, with a network of think tanks and conservative media outlets, the billionaires bankroll the spread of this neoliberal market fundamentalism and pay politicians to do nothing. Also block the kinds of sensible regulations other countries enjoy and the public naturally demands in the absence of all this brainwashing. This is why in the progressive worldview, it is so notable that the billionaires themselves are calling for regulations on AI. It signals to them that this is genuinely a serious issue beyond just about anything else we've encountered so far. And when advocating for the kind of regulations progressives often want, it hands them the talking point that with this industry, even the top CEOs admit we need more regulations. To us opponents of government regulation, they can and do tell us that even the evil billionaires we're supposed to be serving are on board with regulations. So, we're just making a fool of ourselves if we continue opposing them. This entire narrative may sound coherent enough, especially considering how much we're exposed to it by our teachers, textbooks, the news media, and Hollywood, but it is entirely wrong. And that becomes obvious if you just look at history. There was a time when the American economic system could reasonably be defined by a lack of government regulation, but it was well over a century ago in the early to mid-1800s. The country and really the entire world was far poorer than we are today. So living and working conditions were significantly worse on nearly all fronts than what we living in 2026 would deem acceptable. But at the same time, state governments and to a far greater extent the federal government were far weaker than what we now live under. The inability of government officials to get away with intervening heavily in the markets allowed the market process to thrive. So while the poverty of the time and everything that entailed was certainly nothing to envy, wealth was being produced and therefore poverty was disappearing at a pace rarely seen in human history. However, it was not enough for everyone. Several businessmen in various industries who had already achieved some level of market dominance did not like the fact that to maintain their position, they had to constantly compete and offer consumers better goods and services at better prices than any of their competitors and any potential future competitors. As these titans of industry saw it, that kind of cutthroat competition and permanent vulnerability was beneath societal elites like themselves. As Murray Rothbard wrote at the beginning of his book The Progressive Era, these increasingly lazy entrepreneurs made several serious attempts to come together and form nationwide cartels. If all of the big firms in an industry could just agree to not undercut each other on price or unfairly offer their customers a quality that was too high for the agreed upon price, it would immunize the companies from competition and allow them to dictate to consumers the prices and quality of goods they would provide rather than the other way around. But it never really worked. The ease at which a new entrepreneur could enter a cartelized market and undercut everyone made it essentially impossible for any genuine industry-wide cartel to get off the ground. So these market elites that didn't want to compete turned to the government. By the mid to late 1800s, there had already been some so-called political entrepreneurs who had chosen to focus on seeking political favors to secure their wealth and market dominance rather than producing value themselves. But because, again, governments were so small and so localized in the early days of the country, this usually amounted to little more than an easily ignored official right over some local route or resource. But then that began to change. It started with the railroads. As America's first genuinely large-scale business that more often than not expanded across multiple states, it's not surprising that this is where we find the first successful federal government-backed cartel. After decades of trying and failing to form lasting voluntary cartels, the railroad giants began calling for government oversight in the 1870s. That oversight really just boiled down to the government using its power to enforce the cartel for the big railroad companies. It was a deliberate campaign by the rail industry to enrich itself through laws and regulations that it and its intellectual allies presented to the public as limitations on the industry that were purely in the public interest. That effort culminated in the Interstate Commerce Act and the ICC in 1887. And with that, a successful approach for creating cartels, shielding oneself from the need to compete, and reaping lucrative political privileges had been developed. As Rothbard laid out in detail, that approach was adopted and refined throughout the 1890s, 1900s, and 1910s, the period we now call the Progressive Era. Firms in the steel, sugar, heating oil, meat packing, and agricultural industries all began emulating what the railroad companies had done, and that strategy quickly spread from there. The government officials gaining power and perceived importance as a result were, of course, happy to go along with it. And after the Cleveland Democrats collapsed in the election of 1896, political support for the emerging crony rackets became bipartisan. This was truly a revolution in the American economic system. Industry leaders and their allies in government came to understand that they could get away with all kinds of state-enabled extortion of the public as long as they found some way to sell it to that public as a necessary common-sense check on corporate power that was in the public's own interest. And so, for many decades now, heads of industry and their political allies have searched for and cycled through any effective excuse to further suppress competition in their industry and gain additional political privileges. Just look back at the formation of the banking cartel known as the Federal Reserve, the precedent-shattering crony monstrosity known as the New Deal, the health care industry-enriching, price-amplifying programs called Medicare and Medicaid, the precedent of bailing out the financial sector whenever it's in trouble, the massive health insurance industry bailout known as Obamacare, Biden's extensive federal rules to mitigate climate change, and much more. All of these were crony power grabs that helped form de facto government-backed cartels that then used state power to transfer more and more of the public's wealth to the members of those cartels. But they were sold to the public as necessary measures to keep that industry in check and to keep us, the public, financially and physically safe. What we are seeing today with this sudden consensus among the big AI CEOs that they need to be regulated is simply an attempt by the big firms in a young industry to establish the kind of government-enabled cartel that so many other industries enjoy. There is not yet complete agreement about what that looks like, but all the proposals from the Anthropic, OpenAI, and DeepMind CEOs share the same basic elements. All want some kind of US government agency or US government-led international body that can mandate extensive safety testing, set standards, certify compliance, and block new models if the agency determines it's in the public interest. If enacted, that would be the first step towards a textbook 1880 railroad industry style cartel. It has the necessary elements, expensive safety testing and compliance requirements that none of the current leading firms had to invest time, money, or resources in when they were getting started, and an entity with state power that can block business practices it declares unsafe or somehow against the public interest. If these firms succeed in getting the government to bring all this about, their market dominance will be far more secure. Many call this regulatory capture, but that really is a misnomer. Capture implies the regulatory apparatus was established to impose limits on the industry and is only later diverted to serve the interests of the top firms in that industry. But as history makes clear, the regulatory state was built from the beginning with the express purpose of serving the interests of well-established and well-connected companies. In other words, the heads of big companies pushing for more government regulation in their industries is not some strange deviation from the status quo. It's an embodiment of it. If the AI CEOs truly believed that the technology they were working on would kill all humans in the next 3 years, they would be acting differently. Like so many of their predecessors, these executives are trying to scare us into supporting policies designed to protect them from competition. We need to stop falling for it. >> [music] [music]