I recently proposed that a tax on the purchase of human attention might resolve one of the key issues at stake in the ongoing debate about how the government should regulate free speech: the purchase of mass distribution and the conversion of financial capital into social/attention capital and political capital. The proceeds of this proposed tax on the exercise of power would be distributed pro rata to citizens, so that exercise of economic power would directly result in redistribution of that power. Here is the full post:
Speech is Free, Distribution is Not // Revisiting the Attention Tax.
Originally, that post was much longer, and included a lot of historical and present day context as the backdrop for the proposal, but feedback from early readers led me to cut drastically to make a tighter argument there.
You can think of this post as either the extension of or prequel to the proposal for a tax on a purchase of human attention.
NB: if you read both, you may recognize a few small snippets are duplicative, but these are mostly non overlapping…
In attempt to arrive at informed opinion about the debate around the governance of free speech on the internet, I want to explore some of the historical context on the origins and goals of the right to free speech, review some historical debates around restrictions of free speech, and consider the implications of the key differences between past and current forms of communication technology.
While the basic challenges to the maintenance of public order presented by the right to free speech are not new, new technologies have changed the economics of free speech — specifically, technology has dramatically democratized and lowered the cost of production and distribution of content, while increasing the cost of filtering and verifying information (because we have not invented equally powerful new technology to keep pace filtering noise and verifying accuracy of accelerating information production).
Debate Around Section 230
The role of social media on the outcome of the 2016 US presidential election, President Trump’s daily allegations about fake news (and his seeming total disregard for the veracity or consistency of his own public discourse), the decline of journalistic integrity and transformation of the news business into yet another entertainment business, the bannings of controversial figures by internet media platforms… these are just a few of the things that have inspired debate about the governance of free speech on the internet.
Specifically what is under debate is Section 230 of the Communications Decency Act, which both (i) absolves internet platforms from the legal responsibility of policing potentially harmful discourse (like fake news) on their platforms:
No provider or user of an interactive computer service shall be treated as the publisher or speaker of any information provided by another information content provider.
and (ii) gives internet platforms the discretion to moderate content on and ban users of their platforms at will:
No provider or user of an interactive computer service shall be held liable on account of — (A) any action voluntarily taken in good faith to restrict access to or availability of material that the provider or user considers to be obscene, lewd, lascivious, filthy, excessively violent, harassing, or otherwise objectionable, whether or not such material is constitutionally protected….
Advocates of reforming part (i) argue that given the role that internet media platforms seem to have played in influencing the 2016 election (and continue to play in the modern political process more broadly), internet companies should be held responsible for the content distributed on their platforms, more like TV and radio broadcasters.
Advocates of reforming part (ii) argue that when platforms exercise the right to moderate content and ban users for what the platforms deem to be harassment or violent speech, this entails a violation of the banned/moderated users’ own first amendment rights to free speech.
Historical Context
Before digging further into the debate at hand, let’s first consider some historical context.
First, why is free speech a constitutional right at all?
If you think about it, democracy implicitly requires free speech. If citizens are to govern, they need to be able to freely critique the current regime, rulers, and laws. The first amendment specifically cites this “right to petition the Government for a redress of grievances.” And, it protects not only the right to free speech but also to press and assembly, making specific point to protect distribution technology and reach, as speech without distribution and audience is meaningless / powerless to inspire political change.
The right to free speech is not absolute, however. It is not framed as an end in and of itself, but as a tool for achieving a higher end, namely, for critiquing government. So there is a long history of debate about restrictions over free speech.
One area where the government has decided repeatedly that speech is not protected by the first amendment is when “speech is intended to result in a crime, and there is clear and present danger” — in Schenck v. United States the Supreme Court decided distribution of pamphlets promoting draft evasion presented such danger and violated the Espionage Act of 1917, for example.
Many of the restrictions the FCC places on broadcasters involve gradations of this concern (ie, interpretations of what constitutes clear and present danger). Specifically, the FCC restricts the broadcast of Hoaxes, News Distortion, and Programming Inciting “Imminent Lawless Action” (the FCC regulations broadcasting are an interesting, fairly comprehensive read to get a sense of what categories of speech the government limits on public tv / radio and why).
The other major area where the FCC restricts speech and communication technology is Political Broadcasting (ibid). The government recognizes that candidates that have more airtime will have an advantage in elections (and, conversely, no one will vote for the candidate who has never been on tv and they have never heard of). Towards this end, the FCC requires that broadcasters provide reasonable access to all candidates for federal office to purchase airtime, that they provide equal opportunities to all candidates in purchasing airtime, that they charge all candidates the same rates, etc.
Given this acknowledgement that more airtime effectively translates to political advantage, there has been much debate around the political spending — because it is possible to buy human attention in the form of ads, you can trade money for votes by purchasing more ads for your campaign. Allowing unrestricted spending on political campaigns thus gives an unfair advantage to the wealthier candidates competing in an election.
There also has been much political debate around campaign spending more broadly, under the premise that allowing unrestricted spending on political campaigns gives an unfair advantage to the wealthier candidates competing in an election. In an attempt to remediate this concern, the Federal Election Campaign Act of 1971 required disclosures of contributions to federal campaigns and an amendment to it placed limits on campaign contributions and expenditures. Subsequent supreme court cases citing the first amendment, however, found some of these restrictions unconstitutional. Buckley v. Valeo ruled expenditure limits contravened the first amendment because they restricted “quantity of expression,” Citizens United v. FEC held that restrictions on corporate spending on “electioneering communication” were unconstitutional, and McCutcheon v. FEC held that aggregate limits on political giving by an individual are unconstitutional. These rulings all protect the ability to trade financial capital for political power.
Internet Media Platforms vs Broadcasters
At first glance, it is not clear why internet media platforms should not be subject to the same rules and regulations (governing Hoaxes, News Distortion, Programming Inciting “Imminent Lawless Action,” and Political Campaigns) as broadcasters.
Why did Section 230 carve out exceptions for internet platforms?
Sen. Ron Wyden, one of the authors of the law, has said the intention was to allow internet companies to focus their resources on innovation rather than policing content:
There were innovative new businesses sprouting up all over, and novel forms of communication and media connecting and informing people in new ways. But it was clear that the quickest way to strangle that revolution in its infancy was for those new companies to be held legally liable for every piece of content users posted on their platforms…. I wanted new small businesses to start out focusing on hiring engineers, developers, and designers rather than worrying about needing a team of lawyers.
This brings us to the question of what is new and different about our current context and technology and what parallels the historical challenges and governance of communication technology. Let’s consider this question on a few dimensions:
Accessibility, Scale, and Cost of the Internet. Anyone can publish content, effectively for free, on the internet. And, anyone can host / curate / distribute content published by others to their own audience, again at near zero cost. The accessibility of publishing mirrors more closely the creation and distribution of pamphlets, which pretty much anyone could do, but had pretty limited reach. In the past, the ability to reach millions of people was held primarily by large mass media corporations, and the ability for an individual to reach millions of people, at essentially zero cost, is unprecedented.
Social Proof. Advertising and marketing have long leveraged social proof, understanding that I am more likely to trust my tribe than I am to trust mass media or an unknown corporate endorsement. I remember these “Rock the Vote” campaigns on MTV from when I was a kid, which showed you a bunch of kids wearing grungy clothes like you wore, reminding you to vote. These were slick, because they told you just to vote, not who to vote for, but you can bet the people who sponsored this campaign knew exactly which direction you were going to vote. While this kind of social proof marketing has long existed, what is new and different today is the ability to pay to show me my actual friends’ faces as an endorsement of your brand or campaign. This is far more effective than showing me an endorsement from a stock character deemed representative of my tribe. I’ll respond differently to a television advertisement on ABC for a presidential candidate that comes straight from the party vs. a reminder to vote from someone I identify with on MTV (or Vice or Fox News or NPR) vs. a story that one of my actual friends shared on social media about AOC.
Polarization and Filter Bubbles. There is much concern over the contribution of micro-targeting and the algorithmic recommendations and content curation in fueling polarization. While it is easier than ever to target small groups with custom messages, the polarization of media is nothing new and is not a problem specific to internet media — talk radio and cable tv (the publications I mention above, MTV, Vice, Fox News, NPR are all good channel examples, Rush Limbaugh, Howard Stern, The Daily Show, Tucker Carlson are program examples) have long targeted specific groups with pretty one-sided content. I’m not sure the delta in filter bubbles is as significant as the change in ability to pay for social proof.
Optimization at the Content Level (vs the Program or Channel Level). Historically, publishers could measure the popularity of their content based on things like the circulation of their paper, a survey about the viewership of their tv or radio station (or maybe ratings about an individual program on the station). Today, optimization can and does happen at the story or headline level, because this is what you sell ads against, drive traffic to, and expand your audience of social followers through with shares, likes, and retweets. I believe this is the primary driver of what seems to be a shift towards more and more extreme and outraged content on the internet (far more so than the increased ease of locating a niche community / tribe).
Fake News. As we saw in the guidelines the FCC uses to regulate broadcasters, hoaxes have long been a problem with the media. The new thing here is not fake news itself, but the low cost of distributing it, and the ability to show me the faces of people I actually know and trust endorsing fake news.
Some Proposals
Given the specifics of our current situation, what (if anything) should we do to address the challenges of preserving free speech while maintaining social order?
Education. One idea that we will almost certainly not hear much about from the government or mass media — because (i) it is difficult and involves civic responsibility and (ii) it is arguably at odds with the interests of incumbents who understand how to manipulate public onion of and make money on an uneducated public using the current forms of communication technology at their disposal — is arming the public to better grapple with misinformation in critical thinking. Connor Leahy proposes this in his post about the implications of GPT2, and many governments are already adopting this strategy.
Tech for filtering and verifying accurate information. While there has not to date been parallel advancements to combat the new technologies for production and distribution of disinformation, this does not entail that it is impossible to build this complementary tech (although, if we go this route, we should consider the tail risk of outsourcing this task to machines). A slightly skew path that seems promising here is a sort of arms race argument where we use technology raise the bar for what constitutes accurate information — the interesting example of this I came across recently was recording all video from 4 angles simultaneously, and presenting these in a grid, with the idea that it would be far harder to spoof this with deepfakes than a straight on single camera video.
Tax the conversion of financial capital into attention and political capital. I explain the mechanics of how this might work in the proposal which this background context was originally part of:
→ ****Speech is Free, Distribution is Not // Revisiting the Attention Tax**.**
Conclusion
While the existential risk that free speech poses to democracy is nothing new, new technology for producing and distributing speech has changed the economics of speech dramatically enough that we might want to rethink our strategies for mitigating these risks.
A final quote from Alexis de Tocqueville:
I approve of [liberty of the press] more from a recollection of the evils it prevents than from a consideration of the advantages it ensures…..[I]t is at the same time indispensable to the existence of freedom, and nearly incompatible with the maintenance of public order.