On Censorship

At City St George's University this week for a discussion on Truth, Trust and Tricksters in the Age of AI, organized by Index on Censorship, the Institute for Creativity and AI and Global (Dis)Order.

One of the panelists, Kenneth Cukier of the Economist, recounted his attempt to get an AI tool to produce an image to illustrate a point about silo-busting. The tool refused on the grounds that this would be an image of destruction. Cukier regarded this as an example of censorship - the tool was denying his ability to express his ideas freely.

As I suggested in the subsequent discussion, we might also regard this as the tool censoring itself. Self-censorship has been a feature of many authoritarian regimes in the past, and is perhaps emerging in new forms today, especially as people may fear the consequences of their words being taken out of context and blasted around the Internet.

One of the features of agentic AI is presenting the AI device as an autonomous agent. It may appear to be working for Crukier, but it is ultimately controlled by whichever big tech assemblage has developed and disseminated it. Which brings me to a critically important question I have asked several times previously - Whom Does The Chatbot Serve? (Towards Chatbot Ethics, May 2019)

In any case, all communication and creation is necessarily selective. However many images the AI tool does in fact produce, there is a much larger quantity of possible images it has decided or been programmed not to produce. There must be huge amounts of material that the editors of the Economist choose not to publish, for whatever reason, and it would be absurd to frame all these editorial decisions as forms of censorship or self-censorship (although of course some of them may well be).

However, attempts to reduce the quantity and reach of misleading content will often be framed as censorship by those promoting such content, as Jacob Weisberg's latest article demonstrates. There are some difficult issues here, and it is sometimes hard to avoid taking a political side.

 

See also Thinking with the Majority (May 2021), Amplification and Attenuation (October 2021)

Jacob Weisberg, Algorithm Nation (New York Review of Books, 23 October 2025) 

Amplification and Attenuation

@STS_News (Lee Vinsel) poses a couple of important questions on Twitter

• Have any of y'all seen any specific proposals for how to regulate Big Tech companies around mis/dis/harmful information?
• Are there any so-far best philosophical/ethical treatments for why platforms *should* shutdown groups like Qanon?

I think it makes sense to start with the WHY question before tackling the HOW. In my post on The Ethics of Diversion - Tobacco Example (September 2019), I listed four ethical principles that might be relevant to that case, and discussed the need to balance these appropriately.

• Minimum interference principle
• Utilitarianism
• Cautionary principle
• Conflict of interest
  

If the primary justification for shutting down certain groups is to reduce the circulation of ideas that society has agreed to classify as undesirable, for whatever reason, then the minimum interference principle urges us to consider alternative (less extreme) measures that might achieve almost as much suppression with less interference. For example, instead of banning these groups from a given platform, it might be almost as effective to allow these groups to post whatever they like, provided that these posts are not picked up and recommended to others by the platform algorithm. Indeed, it might be safer to keep these groups on a well-controlled platform, rather than push them to migrate onto a less well-controlled platform.

In other words, in order to justify total shutdown, you have to demonstrate that this will produce significantly better outcomes than lighter regulation would.

However, in the case of the social media platforms, this is complicated by conflict of interest. The platform algorithms are designed to promote the commercial interests of the platform, and this means maximizing the engagement (or even addiction) of the so-called users. So society may not feel it can trust the platforms to manage the recommendation algorithms in a way that achieves the outcomes that society wants (and avoids the outcomes that society wants to prevent). And the algorithms themselves are generally not amenable to independent inspection, being incredibly complex and wrapped in commercial secrecy.

Meanwhile, not everyone is comfortable with the idea that ethical decisions should depend solely on the consequences of this or that action. If there is something clearly objectionable in the content that a given group wishes to post, or if we can demonstrate some evil intention behind it, surely we should be allowed to object to the content itself, not merely suppress its consequences? It may also be difficult to determine these consequences with any degree of accuracy, either in advance or retrospectively. So in addition to the consequentialist arguments we have already looked at, moral philosophers have identified a different class of argument, which they call deontological.

This includes arguments based on the intrinsic value of truth, and the idea that platforms should suppress material simply because it contains certain kinds of falsehood. However, I think this idea is insufficient as a basis for regulation, as I argue in my post on Ethical Communication in a Digital Age. 

There's obviously a lot more I could say about the WHY, but let me say a few words about the HOW. One way of understanding regulation and control comes from cybernetics. As we see from the work of Stafford Beer, for example, the behaviour of complex systems can be controlled by a combination of amplification and attenuation. In the world of social media, the critical question is which messages are selected for amplification.

In her article on Amplified Propaganda, Renée DiResta suggests that traditional top-down propaganda (which Chomsky called the Manufacture of Consent) has been replaced by the bottom-up creation and selective amplification of narratives that shape reality. She calls this Ampliganda.

In many cases, these narratives appear to emerge spontaneously from the crowd - this is sometimes called going viral. However, there are increasingly sophisticated ways of engineering and orchestrating these narratives, not just with bots but by convincing real people to participate in their propagation.

Meanwhile, if some things are being amplified, other things are being suppressed or ignored. In most cases, this is not the result of anyone actively deciding to suppress them, more a consequence of the fact that there is a finite quantity of attention. If a book or film or other cultural product receives poor reviews, or perhaps nobody even bothers to review it at all, then it is unlikely to become a best-seller. Similarly, the majority of the content on the internet will be read by at most a handful of people. It doesn't make sense to accuse the platform of exercising censorship, or to complain of being cancelled, simply because something fails to achieve a mass audience.

Furthermore, there are some structural reasons why certain types of content fails to attract an audience. In politics, there are limits to the kinds of idea that people are willing to take seriously - this is known as the Overton window. (However, this can change over time, for reasons that I can't go into here.)

But while the majority of people are comfortably inside the Overton window, there may be a significant minority who are attracted to groups positioned well outside the Overton window. So instead of asking how we should regulate the use of Big Tech by these groups (or some might say the cynical exploitation of these groups by Big Tech), maybe there are some larger questions about the regulation and/or liberation of discourse inside and outside the Overton window.

---

Renée DiResta, It’s Not Misinformation. It’s Amplified Propaganda (Atlantic, 9 October 2021)

Stanford Encyclopedia of Philosophy: Consequentialism, Deontological Ethics

Wikipedia: Overton Window

Related posts: Ethical Communication in a Digital Age (November 2018), Culture War - What Is It Good For? (July 2021)

Culture as a Lens on the Future

In my piece on the New Economics of Manufacturing (November 2015), I mentioned Jacques Attali's idea that culture can provide clues about the future.

For Attali, music is not simply a reflection of culture, but a harbinger of change, an anticipatory abstraction of the shape of things to come. from a review of Attali's 1985 book Noise

I have now been reading about a recent research project in Germany, which used literature, not to predict the future exactly, but to identify potential trouble-spots. The project was led by Jürgen Wertheimer, a professor of comparative literature at the University of Tübingen, and part-funded by the German Ministry of Defence.

The project was able to detect weak signals of sociopolitical and ethnic conflict, not only from contemporary novels but also from the cultural response they generated. For example, by analysing relevant fictional material, they were able to demonstrate worsening sentiment between the Albanian and Serbian communities more than ten years before the Kosovo crisis of 1998. Applying the same technique to Algeria, they picked up weak signals of impending crisis two years before the events of February 2019. They also provided advance warning of conflict in Nagorno-Karabakh.

If verbal conflict provides advance warning of real conflict, can literary action also be used to defuse conflict? Governments clearly think so, intervening with positive propaganda to promote the desired messages, as well as censorship, exile or worse for writers who touch a sensitive nerve. But not only is this intervention often counter-productive, it also helps amplify the warning signals for those such as Professor Wertheimer who are looking out for them.

There is a simplistic view that media and culture can cause social and political events - this is known as the media effects narrative. For this reason, political actors often wish to control media and culture, as a way of managing social and political change. Does the control then become the message?

---

Philipp Oltermann, ‘At first I thought, this is crazy’: the real-life plan to use novels to predict the next war (The Guardian, 26 June 2021)

Studienprojekt Cassandra - not to be confused with the US Project Cassandra (Wikipedia)

The Wrong Side of History

In his inaugural speech yesterday, President Barack Obama spoke the following words: "To those who cling to power through corruption and deceit and the silencing of dissent, know that you are on the wrong side of history." [President Barack Obama's Inaugural Address]

Anyone who thought those words were directed at the Cubans or the Iranians or the Filipinos or Israel or Jimmy Carter or Microsoft would have been alerted to their true target by the swift censorship of the Chinese authorities, which excised these words from the Chinese translation (along with some unfavourable references to communism and blaming the West) [BBC News 21 January 2009].

As BBC Journalist James Reynolds reminds us, President Bill Clinton used similar words before making a "friendship" visit to Beijing in 1998. "When it comes to human rights and religious freedom, China remains on the wrong side of history" [CNN, June 1998]. And in his speech at Beijing University [BBC News, 29 June 1998] he asked his hosts "How do we work together to be on the right side of history together?" Apparently Clinton's views on history were shared by disgraced former Chinese leader Zhao Ziyang [New York Times, 25 June 1998] See also Jonathan Fenby [Guardian, March 2008]

Does history have a right and wrong side at all? Is history some kind of bully, that we have to stay on the right side of or get beaten up? Is there an inevitable march of progress and freedom, or is that just what Butterfield called The Whig Interpretation of History? (See commentary by Dr John Warren). Butterfield charged earlier historians who believed naively in such progress, especially Macaulay, with "fatuous and complacent optimism". Carl Becker levelled similar charges against Jefferson (see David Noble, Historians Against History, p 91).

Pray Obama does not succumb to fatuous and complacent optimism.