High Profile Commission On Disinformation Unable To Solve Disinformation Despite Having Prince Harry On Board
from the i-mean,-you-tried dept
Dealing with disinformation is not an easy problem to solve. Part of the problem is that very few people even agree how to define disinformation, or how subjective it is. Indeed, as we've noted, most of the reporting on disinformation itself is misinformation (or, at the very least misleading). That said, I still had decently high hopes for the Aspen Institute's "Commission on Information Disorder." The Aspen Institute tends to do more credible and serious work on tech policy issues than many other groups. And the project was supported by Craig Newmark, who has been funding a bunch of important research over the past few years. And, while some of the choices for who was on the Commission struck me as odd (Prince Harry?!? Katie Couric?!?), there were some very serious and very thoughtful participants on the Commission itself, acting as "advisors" to the group, and who participated in the various discussions they held.
But, perhaps the wide range of perspectives of people involved was more of a hindrance than a help. The final report that was just released is a kind of punt -- in which the Commission effectively tries to "split the baby" -- by offering a kind of middle of the road perspective, without realizing that splitting the baby was never the real goal of the anecdotal judgment.
That is, the report more or less acknowledges that the real problems are much more fundamental than disinformation (indeed, it quotes me on this very point -- even giving me a pull quote treatment -- though at no point did anyone involved in this project reach out to me or ask me for any input), but then still puts in place recommendations that don't seem to acknowledge this reality. So upfront it admits that there is a fundamental issue in that a society with arbiters of truth is not a free society:
The Commission recognizes that in a free society there are no “arbiters of truth.” As we each then evaluate which sources news consumers, citizens, and tech platforms may rely upon to inform their own decision-making—whether they are professionals, academics, nonpartisan organizations, or government institutions— we must acknowledge that this is part of a much bigger process. The news media have a responsibility to seek out the facts, and report them with relevant context for their audience. In empirically-grounded realms, we can reduce information disorder by enabling and supporting efforts based on transparent, reproducible, proven methods (e.g., scientific method for health, certification methods for elections) implemented by non-partisan, qualified professional organizations, and accountable public oversight (e.g., medical/health organizations, election process orgs).
It also highlights that perhaps "disinformation is a symptom" of a larger problem, rather than the problem itself:
Mis- and disinformation are not the root causes of society’s ills but, rather, expose society’s failures to overcome systemic problems, such as income inequality, racism, and corruption, which can be exploited to promote false information online. As Mike Masnick has written, “Saying that the disinformation is the problem—rather than a way in which the underlying problem shows itself—misses the point entirely.”
Mis- and disinformation do not exist in a vacuum. The spread of false and misleading narratives, the incitement of division and hate, and the erosion of trust have a long history, with corporate, state actor, and political persuasion techniques employed to maintain power and profit, create harm, and/or advance political or ideological goals. Malicious actors use cheap and accessible methods to deliberately spread and amplify harmful information.
It also admits that our belief in free speech means "doing something" about disinformation will likely conflict with the 1st Amendment.
First Amendment scholars and advocates have raised legitimate concerns about undermining free speech as society tries to address the most harmful content in the digital public sphere. Commissioners concluded that introducing new constraints on harmful speech would likely lead to unintentionally censoring “socially valuable speech,” like dissenting views and advocacy for minoritized, underrepresented, or vulnerable people and groups, or protected classes.
Given all of that, you would hope that the report would then focus on looking at ways to alleviate some of those underlying societal issues. But when you get to the recommendations section, it's almost as if it was divorced entirely from the first part of the report, and took very little of that into account. Admittedly, there are a lot of recommendations, and some of them seem perfectly reasonable. But the perfectly reasonable ones are unlikely to have much impact. And the ones that might have an impact seem unlikely to have the impact the authors seem to expect.
Since there are so many recommendations, I'm not going to go through them all, but I will start out with the most concerning one: changing Section 230. To be fair, the changes they suggest are somewhat limited, but still would be hugely problematic (and almost certainly unconstitutional). But, without even getting into why the suggestions are problematic, the media is already taking the "change 230" suggestion and running with it. It's buried in the back of the report, yet this is how Fast Company titled their article on this report:
That says:
Katie Couric, Rashad Robinson, and Chris Krebs say it’s time to pull immunity for social media platforms
And the subhed says:
In a new report on disinformation from the Aspen Institute, experts propose revising Section 230 and forcing social media companies to submit to data audits.
So the takeaway, is "take away 230." Even though that's not actually what the report says -- but it demonstrates how giving an inch on Section 230 means that some will seek to take a mile.
As for the actual "reforms" the report presents for Section 230, there are two separate ones. First, remove Section 230 for "paid content" which is not well defined beyond "advertising" (even though "paid content" can mean a lot more than advertising):
The Commission recommends that paid advertising content should no longer be protected under Section 230. Similar proposals have been advanced by legislators, academics, and advocates, including a bill before the House of Representatives, co-sponsored by Members of Congress Malinowski and Eshoo. Tech platforms should have the same liability for ad content as television networks or newspapers, which would require them to take appropriate steps to ensure that they meet the established standards for paid advertising in other industries. This amendment to Section 230 is tied to paid promotion to amplify or promote content, not all paid content services or subscriptions (e.g., Patreon, Netflix, OnlyFans). It is not intended to impact Section 230 (c)(2), which protects platforms from civil liability when they decide to remove objectionable content.
This part seems entirely at odds with the cautions earlier in the report about messing with 230 and the 1st Amendment implications there. It seems to assume (incorrectly) that those can be ignored when it comes to advertising or paid promotion. While it may be true that commercial speech historically is given less protection than non-commercial speech, that does not mean it has no 1st Amendment protections at all. But, a bigger point is that this kind of bill would be hugely problematic.
The ability to promote things is a key part of how small businesses succeed today. Tons of businesses rely on the ability to advertise through platforms like Google and Facebook, and increasing liability on such things will make that much more difficult. Google revolutionized the ad market by making it easy for any business to reach a very targeted audience, and to do so more or less automatically. That's not to say the system is perfect, but it has created tremendous benefits for many, especially small businesses, and adding liability to the process would greatly limit it for small businesses.
Even worse, at a time when lots of folks are worried about the power and control that Google and Facebook have over this market, a change like this to 230 would lock in those two providers as more or less the only ones that could take on the liability. Thus it would do damage to competition.
Also, it simply remains unsubstantiated that this would fix anything. Most of the content that people are concerned about regarding disinformation isn't being spread through paid advertisements. Sure there are a few odd cases here and there, but there's little evidence to suggest that the spread of disinformation is really driven by paid promotion.
Perhaps even more odd is that in the "rationale" section for why they want this change, the report's authors... never actually explain the rationale. It's quite odd. They just talk about how old Section 230 is and how it hasn't been updated in a while. But what does that have to do with advertisements and paid promotion? It just briefly mentions that paid promotions are "used by bad actors to promote questionable health cures, and by political campaigns to smear opponents."
But... that leaves out an awful lot. (1) Political campaigns have always smeared opponents. That's not new to social media or the internet. It's also (whether you like it or not) protected under the 1st Amendment with only a few extremely limited exceptions (out and out defamation, which has an extremely high bar for a public figure running for office). (2) The report makes no effort to provide any evidence as to how big (or not!) a problem this actually is. Social media sites have actually been quite aggressive in looking to take down "questionable health cures" because they know those are problematic. Indeed, it's Section 230's broad protections that allow the companies to feel confident that they can moderate this way without fear of legal liability.
So there's no clear evidence that paid promotion is the problem. There's no clear evidence that changing 230 over this issue would "fix" any of the so-called problem, considering that most of it is protected under the 1st Amendment anyway, and thus any lawsuits would fail. So how is this possibly a thoughtful and considered recommendation?
The second change to Section 230 would be to remove it for recommendation systems:
A Section 230 exclusion for product design features would clearly delineate between the speech of platform users and the platforms themselves. This would include explicitly designating the output of recommendation algorithms (group recommendations, friend recommendations, “watch next” recommendations) as not being user speech and therefore not covered by Section 230 (to the extent that the recommendations themselves might cause harm).
Again, this has been suggested by others, and would have massive 1st Amendment implications. A recommendation is speech. A recommendation is an opinion. And that's protected under the 1st Amendment. Removing it would discourage sites from offering their recommendations and thus would be a move by the government to suppress speech. That's a pretty big 1st Amendment problem. Clearly some of the people this Commission spoke to pointed this out -- as explained in the first part of the report -- but when the authors got here, they appeared to have forgotten this pretty key point.
The rationale section here is even worse -- as it takes a misunderstanding of what happened in two cases and extrapolates out in very dangerous ways:
One clear need is to carve out actions by the platform that lead to real-world harm, but that would fall outside of speech from Section 230. For example, in Herrick v. Grindr, the product tools allowed a malicious user to wage a campaign of impersonation and harassment against their ex-partner. Grindr first ignored requests for intervention, then claimed Section 230 immunity.
In the decision for Force v. Facebook, a case where Facebook’s algorithms were implicated in the promotion of terrorist groups, but ultimately were shielded by Section 230 immunity, Judge Katzman wrote, “As today’s decision also illustrates the extensive immunity that the current formulation of the CDA already extends to social media companies for activities that were undreamt of in 1996. It therefore may be time for Congress to reconsider the scope of Section 230.”
Here's the thing, both cases were clearly decided correctly, even as some push a propaganda campaign (largely driven out of animosity towards Section 230) to argue otherwise. In Grindr, the plaintiffs were mad at the wrong party. Grindr wasn't the one doing the bad stuff. It was the user. And what is always left out of these discussions (in fact, sometimes completely mischaracterized) is that the police arrested the ex-partner and he was sentenced to jail for his actions.
In other words, the criminal system worked. It did what it was supposed to do.
The fact that the guy couldn't blame the platform for what one of its users did wasn't a bug. It's exactly what the law is supposed to do.
As for Force v. Facebook, that was a ridiculous case in which the families of some Americans who were killed by Hamas in Israel wanted to sue Facebook because some Hamas members had, at times, used Facebook. What did the attacks have to do with Facebook? Absolutely nothing. The plaintiffs put forth a totally unsubstantiated theory that Hamas used Facebook to encourage attacks, and therefore the attacks could be blamed on Facebook. That would (maybe?) be more convincing if Hamas didn't have a long history of similar attacks dating back long before Facebook even existed.
There was simply no connection between the attacks and Facebook, and all Section 230 did here was to help get a bad case dismissed with less hassle. The same thing that happened in the Grindr case.
This is the point of Section 230: to stop bogus, costly cases that are mere money grabs, blaming not the party who did anything wrong, but a random 3rd party with a bigger bank account. It's ridiculous that this report would cite those two cases as evidence that 230 needs reform. Those two cases are evidence of why Section 230 is so necessary.
Incredibly, the report claims that this feature -- getting rid of obviously bogus cases at an earlier stage -- is somehow a bug of Section 230, and that it should be changed to allow plaintiffs to go on fishing expeditions for some sort of evidence to prove their claims. That's a recipe for a dangerous level of nuisance lawsuits, encouraging anyone to sue and sift through tons of records looking for anything they can try to pin liability on. This will be stunningly costly for lots of small companies who rely on Section 230. It's a stunning misunderstanding of how Section 230 works, and really discredits this entire report.
Given how much effort and resources went into researching and creating this report, that this is the conclusion they come to is not just upsetting, it raises serious questions about the entire process.
Filed Under: disinformation, katie couric, prince harry, section 230, societal problems
