Who is correct about content moderation in public heath?

Elon Musk or Yann LeCun

Recently, AI researcher Yann LeCun and Tesla/X/SpaceX CEO Elon Musk had a public disagreement. Among his criticism of Musk, Yann is concerned that Elon doesn’t want to restrict speech, specifically on matters of public health.

“This has become particularly concerning since he bought himself a platform to disseminate his dangerous political opinions, conspiracy theories, and hype. He has been quite naïve about the difficulties of running a social network and the (legal) necessity of doing content moderation. One can claim to be a 1st Amendment absolutist, but a lot of content *must* be taken down by law, e.g. terrorist propaganda, child exploitation, blatant hate speech (in the EU and other regions). Then, there is dangerous disinformation that puts public health in danger or corrupts the democratic process. You have to moderate that too. Content moderation is a complicated problem whose best answer is not an attitude of total laissez-faire but a complex trade-off.”

This is a common rallying cry. If we don’t have content moderation we will have, as Yann puts it, “dangerous disinformation that puts public health in danger.”

Is Yann correct that social media platforms should police information on matters of public health?

Consider that there are two errors companies can make. They can fail to remove information that is actually wrong, and they can remove information that is actually correct. Consider also that information comes from 3 places— central authorities (CDC, NIH, Fauci), and individual experts (Jay Bhattacharya), and fringe groups.

The problem with content moderation is not only that we have seen both types of errors, but that the moderators have different rules for different speakers. Central authorities are never censored, and individual experts are. Even as central authorities spout unproven, false, misleading and criminal statements, and individual experts tell the truth. Rarely, fringe groups are correct and censored.

A few examples:

Facebook censored posts that even suggested lab leak might be possible, and now the House subcommittee has unearthed what appears to be a criminal conspiracy to mislead the public about lab leak. Years after this censorship this is currently the NY Times headline— a complete reversal.

Additionally, the CDC has repeatedly and falsely stated that cloth masking slows the spread of COVID, that masking kids < 5 is not harmful, that boosters benefit people who have had covid, that myocarditis from COVID is worse than from the vaccine, and that lockdowns did more harm than good. On most of these issues, the current prevailing view in the media is still false.

Central authorities were never censored on any platform despite countless false statements.

Meanwhile individual experts were repeatedly censored or de-throttled for criticism of lockdowns, masking, vaccine mandates, myocarditis risk for young men, and vaccinating people who already had and recovered from COVID. Martin Kullsdorf had tweets removed and was banned from LinkedIn. These individual experts were broadly correct— but worse, curtailing their speech created the false illusion of consensus.

Now, consider the fringe elements. There have been billions of completely false statements put for by third parties. For e.g. that vaccines contain microchips or that covid vaccines killed all subgroups (e.g. they were net harmful to a 80 year old who did not previously have covid).

Despite content moderation, billions of false claims were NOT removed from platforms. It is hard to know if removing some-- a tiny fraction of these claims— is better than removing none. I haven’t seen evidence to support the claim that removing some false ideas— at great cost and at risk of removing true ideas— actually helped any human being. There is a handful of people in whom these ideas gain traction, and they often seem validated by the mere act of removal.

Putting this all together, I think Yann is wrong, Elon is right, and the pandemic actual exposes the greatest risk: Censorship is not deployed based on the truthfulness of claims, but rather to centralize power and authority and squash dissent. The problem is when the dissenters are correct, in begging us to stop the car, while the central authorities drive us off over a cliff.