LTP News Sharing:
Concerns about ChatGPT – which is an amazing new artificial-intelligence program that can write better than the average Buzzfeed writer and perhaps (a harder test) the average undergraduate (unless it’s just a mirage) – sparked concern among both writers, who realized that they were not immune from work-changing technology advancement, and teachers, who now face an even more daunting task in verifying that their students are really doing the work that they turn in. (A professor friend joked last week that the tell that AI had been involved was the absence of basic grammatical or syntactical errors.)
Quickly thereafter arose concerns about ChatGPT’s by-now-wearyingly predictable bias. It refused to write an argument, for instance, that the continued use of carbon energy is good for human happiness, and consistently takes the left-wing side on hot-button social issues.
As a conservative-libertarian writer (I’ll have to leave it to my readers to decide whether it’s fair to add “and thinker”), my first response to that news was elation. That sort of bias is a full-employment-for-conservative-writers boon. And so long as AI isn’t available for right-of-center arguments, there will be a series of additional benefits. People of the right will still be required to think through their arguments, making them more knowledgeable and engaged. Being a conservative will take more effort and focused discernment than being of more coddled viewpoints, which may to some degree raise the quality of conservatives. Teachers will have more confidence that arguments by their conservative-students are being honestly produced.
The thing is, though, that unless the market for text-generating AI is being artificially constrained in some fundamental sense, this lacuna will not survive for very long. Obviously there is already significant professional and educational demand for the service, however irregular the latter use, and there is sure to be more. The demand for neutral or center/right-friendly versions will be great enough that if market entry is possible, then either ChatGPT’s owners will correct its programming (as they have already indicated an intention to do) or competitors that offer such services will enter the market. Or both. (And anyway, as the editor of these pages wrote in National Review earlier this week, this narrow thinking about technological advancements is always wrong. The development of a ChatGPT-like program that does not exclude center/right subjects or positions will be a boon for such writers, too. What these programs mean is that everyone with access to them will be able to spend less time doing rote work and more time doing more fulfilling work that achieves more, even if it’s different work than they were doing before.)
If neither of those things happen, though, we will have a pretty good idea that artificial market control is being exerted, either by government or by a consortium of competitors and allied market actors. This fairly definitely happened when Apple, Google, Amazon Web Services and others combined to lock Twitter-competitor Parler out of the social-media market in early 2021. That’s illegal: a “combination in restraint of trade.” And corporate executives have gone to jail for much less blatant instances of it in the past.
The ChatGPT-bias controversy, then, provides an object lesson for those grappling with how to deal with efforts by giant fund managers and corporate executives to dictate behavior and eliminate types of competition they personally disfavor without in any way supporting Lina Khan’s “break up all the things” theory of anti-trust that currently controls the FTC (and is getting brushed aside by the courts).
The distinction that the ChatGPT case illustrates is this: Company size doesn’t matter, as the Bork Doctrine rightly recognized decades ago. What matters (in what I think is a corollary of the doctrine rather than an opposition to it) is whether offerings for some legal goods or services that logically should appear in a market, and that there is a demand for, are being by some forces kept from the market or rendered artificially expensive. Where those gaps appear, any other market players that have acted to cause the gap have combined to restrain trade, and should be held at very least civilly liable.
An FTC that sought to address real violation of anti-trust law would therefore have Amazon, Google and Apple – and their CEOs and responsible executives – in the dock for their behavior in 2021. It would not be stopping mergers merely because those mergers “threaten” to offer consumers more goods and services more cheaply and effectively. That leads to the ‘30s-style absurdity of government intervening to keep prices artificially high.
The on-going DirectTV/Newsmax controversy adds additional context. Some months ago, DirectTV (owned by AT&T) dropped conservative news network One America Network. Last week it dropped Newsmax as well. It first claimed that it had dropped Newsmax because Newsmax asked to be paid a carriage fee, which DirectTV does not do. Then it emerged that DirectTV does pay such fees to all the liberal news networks it carries, and that refusing to pay Newsmax was its deviation from normal policy. To objections of what by now had pretty clearly become a pattern and practice of bias by AT&T against conservatives, an unlovely DirectTV executive declared that “[i]f Newsmax ceases to authorize our carriage of their channel [because we refuse to pay it the same way as we do other channels], our customers will still have access to their clearly preferred conservative news channel, Fox News, which has more viewers than MSNBC, CNN and Newsmax combined.”
That comment clarified: This is bias. Note that it has not been AT&T/DirectTV’s policy or practice to refuse to pay, say, MSNBC a carriage fee, because, “after all, liberal viewers still have CNN, CBS, NBC, ABC, PBS, the BBC” and many more choices. For the same reason, it makes no difference that DirectTV has afterward, in a panic, picked up some different conservative news network that no one has heard of. AT&T does not pick-and-choose liberal news outlets, nor does it limit their number or refuse to pay them carriage fees. To do so for conservative networks alone is plain discrimination.
So this is definitely partisan bias, but it certainly does not raise any possible anti-trust issue. There’s at least one more satellite provider, for starters, and satellite TV competes directly with cable and near-directly with streaming apps and Newsmax’ own website. AT&Ts and DirectTV’s channel line-up decisions do not by themselves constitute restraint of trade.
That doesn’t mean that AT&T faces no liability, though. States may want to investigate – to explode the company’s claims that anything but partisan animus drove its considerations. In the course of those investigations, they may want to demand to see the research on which AT&T relied in determining that its bottom line was enhanced by refusing to pay carriage fees to any but hand-picked conservative news networks, thereby reducing its attractions to whole swathes of viewers, while that same analysis and restrictions did not apply to any left-wing outlets. If these states find, as they surely will, that AT&T either has no such research or that the research is garbage, they might want to pass legislation that bars common-carrier platforms from discriminating on the basis of viewpoint, gives customers a cause of action for violation of the statute, and puts the burden of proof on carriers to prove that they’d acted in the objectively established bottom-line interest of the company once a prima facie case of such discrimination has been shown.
Similarly, shareholders might want to demand the relevant records of AT&T and DirectTV to find out the same thing. And should they find an absence of any plausible or objectively established business case (and so find self-dealing, partisan behavior by executives) or one built negligently on rubbish research, they should then bring suit against the corporation. Recall that a finding of self-dealing results in payments coming directly from the relevant executives’ own pockets, and, in the normal course of things, should result in their dismissal.
That would focus some executive minds, I suspect. As, I hope, will the first “combination in restraint of trade” case against, perhaps, some giant investment-house and bank CEOs.
The wokification of American corporate life is no kind of capitalism at all, and great swathes of it is already illegal and can be made more so, and more transparently so, by narrow and constrained state action. In responding to the fact that too many of the hands on the levers of corporate and investment power in the United States have abandoned the law and their duty to shareholders in pursuit of power, we who favor free markets have to take care not to sign up for things that would end up harming those markets, or that freedom, later on. But we also have to move.
Scott Shepard is a fellow at the National Center for Public Policy Research and Director of its Free Enterprise Project. This first appeared at RealClearMarkets.
Author: Scott Shepard