Tuesday, November 24, 2020

Automation, Work and the Achievement Gap

I just published a new paper with my colleague and friend Sven Nyholm. It's in the new journal AI and Ethics and it is about robots undermining human achievements and what we should do about this. One reason why we think this topic is interesting is because it identifies a positive responsibility gap that might be created by autonomous technologies. This is distinct from the negative responsibility gaps already widely discussed in the literature. The paper is available in open access format from the publishers. Links and details below.

Title: Automation, Work and the Achievement Gap

Links: Official; Philpapers; Researchgate; Academia 

Abstract: Rapid advances in AI-based automation have led to a number of existential and economic concerns. In particular, as automating technologies develop enhanced competency, they seem to threaten the values associated with meaningful work. In this article, we focus on one such value: the value of achievement. We argue that achievement is a key part of what makes work meaningful and that advances in AI and automation give rise to a number achievement gaps in the workplace. This could limit people’s ability to participate in meaningful forms of work. Achievement gaps are interesting, in part, because they are the inverse of the (negative) responsibility gaps already widely discussed in the literature on AI ethics. Having described and explained the problem of achievement gaps, the article concludes by identifying four possible policy responses to the problem. 



Monday, November 16, 2020

Wedding Cakes and Discrimination: An Analysis

For some reason, cakes became a major flashpoint in anti-discrimination law a few years ago. In the Masterpiece Cakeshop case in the US Supreme Court and the Asher’s Bakery case in the UK Supreme Court, courts were confronted with similar, albeit not identical dilemmas. In the Masterpiece Cakeshop case, two same-sex customers went into a bakery looking to purchase a wedding cake to celebrate their marriage. In the Asher’s Bakery case, a homosexual rights ‘activist’ looked to purchase a bespoke cake bearing the slogan ‘Support same-sex marriage’.

In both cases, the bakers refused to supply the cake due to their Christian beliefs. In both cases, the respective Supreme Courts held that this refusal was legally permissible though for different reasons. The US decision was a narrow one that overturned a decision by a state tribunal on the grounds that it displayed religious hostility to the owner of the Masterpiece Cakeshop. The UK decision was a broader one, focusing on the freedom of conscience and expression of the baker.

How should we think about these cases? The philosopher John Corvino has written an interesting analysis of the arguments in the Masterpiece Cakeshop case. Looking closely at two of the concurring judgments in the case, one from Justice Kagan and one from Justice Gorsuch, he suggests that there were legitimate grounds for thinking that the refusal of service by the owner of the Masterpiece Cakeshop was discriminatory. Corvino argues that there is an important distinction to be drawn between the use and design of a cake (or, indeed, any object) when it comes to the application of anti-discrimination law: it’s okay to refuse to sell a particular design of cake but not okay to refuse to sell a particular cake based on its possible use. It seems that the UK Supreme Court, agreed with this perspective when reaching their decision.

In what follows, I want to explain Corvino’s reasoning. In doing so, I’m focusing on the general philosophical issues pertaining to discrimination and not on the legal niceties of constitutional interpretation and so forth.

1. Brief Notes on Discrimination

There is no universally agreed upon definition of discrimination (then again, is there a universally agreed upon definition of any philosophically contentious concept?). Very generally, to discriminate between two people, X and Y, is to treat them differently for some reason. If you have two children and you give one of them an ice-cream for dinner and the other one a bowl of cold rice, you are discriminating between them. Some forms of discrimination are justifiable — e.g. it makes sense to give healthcare workers prioritised access to flu vaccines — and some are not — e.g. you cannot fire a woman from her job just because she takes maternity leave. The debate in law and philosophy centres on the dividing line between justifiable and unjustifiable forms of discrimination.

How should we draw this line? Most legal systems around the world do this in a similar way. They define people in terms of different characteristics that they possess (height, age, weight, ethnicity, gender etc) and then distinguish between protected characteristics and non-protected characteristics. Protected characteristics include things like age, religion, race, gender, ethnicity, and, in many places, sexual orientation. It is against the law to discriminate against someone on the basis of a protected characteristic. Non-protected characteristics are not like that. They include things like educational achievement, career history, height, and so on. It is is usually permissible to discriminate between people on the basis of non-protected characteristic (as well as other considerations not pertaining to people). That said, some caution should be expressed because non-protected characteristics are tricky: sometimes they correlate with or stand in for protected characteristics and so can be impugned on the grounds that they indirectly sustain discrimination on the basis of protected characteristics. For example, discriminating against women in employment decisions because they have taken ‘career breaks’ might sound innocuous at a first glance until you realise that ‘career breaks’ can be a proxy for ‘maternity leave’.

Another point to be made here is that sometimes the relationships between personal characteristics and physical actions (the things people do) can be contentious in discrimination cases. Are the things that people do as a result of their personal characteristics integral to those characteristics or not? For example, there is a common slogan among religious believers, in relation to homosexuality, that they can ‘love the sinner’ (the homosexual person) but ‘hate the sin’ (the homosexual activity the person engages in). This can lead them to claim that they are not discriminating against homosexuals when treating them differently or denying them certain entitlements. On the contrary, they will argue that they love and care for them. They just discriminate against what they do. The problem with this line of argument is that it seems a little disingenuous: in discriminating against what they do you are, in essence, discriminating against this specific attribute of their personhood. Maybe you love and care for them as a human in general, but not as a homosexual in particular.

Furthermore, this approach to the relationship between actions and persons gives rise to problems in other contexts. What if we discriminated against Christians for attending church? Could we argue that we are not discriminating against them qua Christian believers but only against what they do as a result of this? This seems implausible. That said, there may be some kinds of activities that are only loosely or indirectly attributable to a protected characteristic, that it is okay to discriminate on the basis of. This distinction — between protected characteristics and the activities associated with them — becomes relevant later on in this discussion.

What justifies the distinction between protected and non-protected characteristics? It’s hard to say for sure. Oftentimes this isn’t well theorised. Usually, the assumption is that protected characteristics are somehow fixed or natural properties of people that are beyond their control. The reasoning then is that it wouldn’t be fair to distinguish between people on the basis of things outside their control. But this argument is not without its flaws and may not explain all cases of anti-discrimination law. For example, people can and do change their religious beliefs all the time. Nevertheless, religious belief is usually a protected characteristic.

Whatever the ultimate reason may be, the issue in the bakery cases is the same. In both cases, the bakers refused to sell items to same-sex people. The argument in both cases that this amounted to discrimination on the basis of sexual orientation. The difficulty in both cases is that this argument seems to butt up against the rights of religious people to freedom of expression/conscience.

Let’s now consider Corvino’s analysis of the cases.

2. Cakes versus Users of Cakes

To understand the arguments in the Masterpiece Cakeshop case, we need to add some more factual and contextual detail, not covered in the introduction. After the facts that gave rise to the Masterpiece case arose, a man called William Jack went to another bakery in Colorado (the same state where the Masterpiece Cakeshop was located) called the Azucar Bakery and asked for a bespoke cake to be made for him. This cake was to be in the shape of a bible, with two grooms on it, a red ‘X’ over them, and an inscription reciting some biblical verses disapproving of gay marriage. Unsurprisingly, the Azucar Bakery refused to make the cake on the basis that they disapproved of the message.

William Jack was trying to prove a point. If it was okay for the Azucar Bakery to refuse to bake the bible-quoting cake, why was it not okay for the Masterpiece Cakeshop to refuse to supply a wedding cake for a same-sex marriage? In both scenarios, the bakers were exercising their freedom of conscience rights not to sell a cake of whose message they disapproved. The only difference — if there was one — is that mainstream opinion was more supportive of the conscience being expressed in the Azucar case. That shouldn’t be a relevant factor. The legal system should protect freedom of religious conscience too, even if many people disapprove of the underlying beliefs.

But is the analogy between the Azucar Bakery and Masterpiece Cakeshop cases as simple and direct as William Jack would have us believe? As best we can tell, there was no proposed slogan or other specifically homosexual symbolic material on the wedding cake in that case. The couple just asked for a wedding cake. Maybe they would have wanted to add some symbols to it if the baker had been willing to sell to them. The problem is that the negotiation never got that far. They asked for a wedding cake and were refused on the grounds that the cake was being used to celebrate a same-sex wedding.

This undermines the analogy that William Jack is trying to draw. The baker in the Masterpiece case doesn’t seem to be disapproving of the message associated with the wedding cake, per se but, rather, the use to which it will be put: the celebration of the same-sex wedding. This looks like it could be grounds for discrimination in a way that denying William Jack the bible-quoting cake is not.

Corvino finesses this point in the following way. He argues that there is an important distinction to be drawn between two things:


Design-Based Objection to a Cake: Objecting to the sale of the cake based on some intrinsic properties it has in virtue of its design.


Use-Based Objection to a Cake: Objecting to the sale of the cake on the basis of the use to which it is to be put (i.e. on the basis some extrinsic properties acquired by the cake in a certain context).


In the Azucar Bakery case, the baker had a design-based objection to creating the anti-homosexual cake. They did not object to William Jack as a customer or a religious believer. They were willing to sell him other cakes. They just wouldn’t make that particular cake, bearing that particular message. Requiring them to do so in the interests of anti-discrimination law would have been to impinge on their freedom of conscience and expression rights.

In the Masterpiece Cakeshop case, the baker had a use-based objection to selling a same-sex couple a wedding cake. He didn’t want to sell them a generic wedding cake because he didn’t like how they intended to use it. Now, to be fair, he was willing to sell them other cakes, and so had no problem selling items to homosexuals in general. He just objected to their symbolic use of the wedding cake. Nevertheless, this seems more like discrimination against those customers on the basis of who they were and not on the basis of what the cake itself said. Requiring him to sell them the cake would not infringe on his freedom of conscience or expression rights. He already made wedding cakes. He was still free to make and create whichever cakes he liked (bearing whichever messages he liked). He just couldn’t control how people used them.

This line of reasoning — that there is an important distinction between the design of a cake and the use of a cake — is essentially what motivated the UK Supreme Court to find in favour of Asher’s Bakery in their version of the cake case. Asher’s Bakery, unlike Masterpiece Cakeshop, were being asked to make a cake bearing a particular message (i.e. having a particular design). Forcing them to make that cake would impinge on their freedom of conscience and expression rights.

3. Objections and Replies

Is this a strong argument? Corvino considers three main objections to it in his article. I want to address these now and offer some of my own thoughts.

The first objection has to do with the whole idea of use-based objections. Corvino’s claim seems to be that in discriminating against the use to which the cake might be put, the owner of the Masterpiece Cakeshop was, in essence, discriminating against the same-sex couple for who they were, i.e. on the grounds of their sexual orientation. It is supposed to then follow that he wasn’t entitled to do this since sexual orientation was a protected characteristic. But is that right? Is there not, as was pointed out earlier on, a distinction to be made between who a person is and what a person does (remember the earlier comments about ‘loving the sinner’ and ‘hating the sin’)?

Corvino makes two points in response to this. First, he argues that oftentimes the way in which we describe the use of something implicates who a person is and not just what they do. This means that objecting to what a person does could be semantically equivalent to objecting to who they are. You could, for example, argue that the owner of the Masterpiece Cakeshop was not willing to sell cakes to ‘people intending to use them to celebrate same-sex weddings’. This is subtly different from saying that he objected to a use to which the cakes might be put. Describing the objection in this way might cause us to question the exact object of his discriminatory attitude. This bit of linguistic trickery, however, is unlikely to persuade anyone. The second, and more telling point made by Corvino, is about ‘identity-constituting practices’. These are activities that are so integral to a protected person’s identity that discriminating against the practice is the same thing as discriminating against the person. As he puts it:

…some practices are constitutive of some identities, such that discrimination against the practice is tantamount to discrimination against the group bearing the identity. 
(Corvino 2018, 12)

This sounds about right to me. But accepting this idea does give rise to some further problems. We could ask the question: is getting married an identity-forming practice for same-sex people? Would objecting to that be tantamount to objecting to who they are? It is at least plausible to suggest that it is not. We could argue that homosexuals can be happy, flourishing and self-endorsing without getting married. (The same is true, of course, for heterosexuals.) That said, the riposte to this is that, in our world, marriage does have an important symbolic meaning and many people will think that it is integral to who they are (and their sense of belonging in a society) to have the right to engage in that practice. In other words, there are some people that will constitute themselves through the symbolic act of marriage. So perhaps it is identity-constituting for some homosexuals.

The only issue with this argument is that it might cause us to wonder whether disapproving of homosexuality is integral to being a Christian, at least as some people conceive of what it means to be a Christian. In other words, if we accept this reasoning is it not then plausible to say that denying William Jack the cake with the bible quotes is tantamount to objecting to his self-constitution as a Christian? Corvino’s response to that is just to fall back on the design/use distinction: no one is objecting to William Jack’s sense of Christian identity or denying him the right to disapprove of homosexuality; they are objecting to making a particular design of cake. But I’m not sure if that is a good response since denying the couple the cake in the Masterpiece Cakeshop case doesn’t prevent them from getting married. It just prevents them from celebrating it with a cake from that particular cakeshop. That said, when we combine this point with the point that Corvino makes in response to the third objection (discussed below) we might be able to resist that conclusion.

The second objection that Corvino discusses has to do with the meaning of the cake. The design-based objection hinges on claims about the meaning associated with particular cake designs. But could you not argue that meaning always depends on context? What a particular cake design means depends, to at least some extent, on the use to which it is put? A generic wedding cake means something different at a same-sex wedding than it does at a heterosexual wedding: in the former case in symbolises an endorsement of same-sex weddings and in the latter case it does not.

This objection has the virtue of being based on something that is true: meaning does change with context. But, as Corvino argues, this shouldn’t affect our judgment about whether a particular design of cake can be rejected or not. A person selling an item cannot possibly hope to control all the possible contexts in which it can be used or the meaning it might take on in those different contexts. To give an example, the person (or persons) that designed the original Guy Fawkes mask cannot possibly be held responsible for, or be allowed to control, the meaning that this mask design has taken on in social protest movements and hacker culture. They can only control the intrinsic features of its design. Ditto for a baker. His or her freedom of conscience covers the right to control the intrinsic features of its design and not all the meanings that might attach to that design down the line.

The third objection that Corvino discusses is the so-called ‘Goldilocks Objection’. This one comes from the concurring judgments in the Masterpiece Cakeshop case in the US Supreme Court. In her concurring judgment, Justice Kagan essentially endorsed the reasoning outline above: that the owner of the Masterpiece Cakeshop may have been impugned for discrimination if we accept that he was objecting the use of the cake and not simply the design. She didn’t phrase it in quite this way but that seems to be the conclusion she reached. Justice Gorsuch objected to this way of looking at it insofar as it was too convenient. To reach the conclusion that the objection was to the use of the cake and not its intrinsic design, we had to describe the cake at just the right level of generality. But what would justify describing it at that “just right” level of generality? As he put it himself:

At its most general level, the cake at issue in Mr. Phillips’s case was just a mixture of flour and eggs; at its most specific level, it was a cake celebrating the same-sex wedding of Mr. Craig and Mr. Mullins. We are told here, however, to apply a sort of Goldilocks rule: describing the cake by its ingredients is too general; understanding it as celebrating a same-sex wedding is too specific; but regarding it as a generic wedding cake is just right. 
(Quote taken from Corvino 2018, 13)


Let’s make this point even more explicit. Gorsuch is saying that the cake at issue in the Masterpiece case can be described in different ways. You can think of it as [just a cake] with no properties unique to weddings. You can think of it as a [wedding cake], intended to be used to celebrate weddings. Or you can think of it as [same-sex wedding cake], intended to be used to celebrate a specific same-sex marriage. The problem with Kagan’s conclusion, in Gorsuch’s eyes, is that it arbitrarily picks the middle-level description, which just happens to support Kagan’s preferred view.

You can imagine what Corvino’s response to this is. He argues that Kagan has not arbitrarily picked the middle level description. She has picked the level that corresponds with the intrinsic, design-based features of the cake: it was designed as a cake to celebrate weddings. She has discounted the lower level description that focuses on the extrinsic properties that a cake of that design might acquire if it is used in a same-sex wedding celebration. If you allowed extrinsic properties of cakes (or other items) to determine their character for the purposes of discrimination law, you would make a mockery of those laws. Any item can be used to celebrate or support a lifestyle or identity under the right circumstance. For example (and this comes from Corvino), birthday cakes are used to celebrate the ongoing life of someone. What if that person is gay, Black or Muslim? Surely it would be discriminatory to not sell a birthday cake to a Muslim simply because it celebrated their life (which you might object to for other reasons)? To make the law workable, you have to focus on the intrinsic properties of the cake, not the extrinsic properties it might take on. This point applies more generally to the sale of other items.

This brings us to the end of this analysis of Corvino’s argument. Overall, I think that he makes some good points. There is some plausibility to the distinction between the design of something and the use to which it is put. Discrimination that results from an unwillingness to supply a particular design of an item might be justifiable; but discrimination that results from an unwillingness to supply an item because it risks being used by people in ‘identity-constituting’ practices of which you disapprove is not. That said, there are limits to this analysis. What counts as an identity-constituting practice might be open to debate. And there are, presumably, other moral limits that might apply to the supply of goods and services that could trump concerns about discrimination. For example, an unwillingness to supply a knife for fear that it might be used in a ritual honour killing, might be permissible. That, however, is an argument for another day.

Tuesday, November 10, 2020

Deepfakes and Sexual Fantasies: Are they both impermissible?

Deepfakes are a type of synthetic or artificial media. By training AI on datasets of images, deepfake technology allows people to create photorealistic fake audiovisual materials. Sometimes these videos depict real people; sometimes they depict artificial people. Deepfakes have provoked philosophical interest in recent times, in part because of the challenge they seem to pose to social epistemic practices, and also because of the significant ethical issues they raise. Is it permissible to create a deepfake video of a real person? If not, why not? If so, under what conditions is it permissible?

These ethical issues are particularly important given the main use of deepfakes. According to one widely-cited Dutch study, 96% of all deepfakes are pornographic in nature. Deepfake technology is commonly used to create videos depicting famous actresses performing sex scenes. Some of the actresses that have been subjected to this include Gal Gadot, Scarlett Johansson and Jennifer Lawrence among many others. Deepfakes are also widely used to create revenge porn depicting ex-partners in sexually compromising positions. Some types of deepfake porn have even been used in political debates and scandals. For example, there have been famous cases of deepfakes (or suspected deepfakes) being used to undermine political opponents in Malaysia and India.

When people hear about these cases, their reaction is often one of intuitive disapproval. They think there is something deeply wrong about the creation of involuntary deepfake porn. If a person has neither consented to have sexual imagery of themselves recorded, nor voluntarily participated in videotaped sexual activity, what could possibly justify or permit their image being involuntarily used to create fake imagery? Surely nothing?

Philosophers, however, sometimes like to ask obvious questions like this to see if they raise any dilemmas in our moral thinking. In his paper ‘Introducing the Pervert’s Dilemma’ (an unfortunate title for reasons I will get to later), Carl Öhman does this for deepfake pornography. In essence, he suggests that if we think the creation of deepfake pornography is morally impermissible, then we should also question the permissibility of generic sexual fantasies. This raises something he calls the ‘Pervert’s dilemma’. Öhman suggests a way out of this dilemma by using the concept of a ‘level of abstraction’ (LoA) to assess different moral phenomena.

Öhman’s paper is a provocative one, and his use of levels of abstraction (LoAs) to resolve the dilemma is certainly worthy of greater consideration, but I’m not sure I agree with key parts of his analysis. I want to explain why in what follows.

1. What is the Dilemma?

The starting point for Öhman’s article is his articulation of the pervert’s dilemma. Here is the basic idea.

Most people have sexual fantasies. In these fantasies, they imagine other people, usually real people that they know or have seen, in sexually explicit scenes. Sometimes they imagine that they are having sex with the object of their fantasy, and sometimes they use these sexual fantasies for self-stimulation. For the most part, we think of these sexual fantasies as a normal and permissible part of human life.

That said, not all sexual fantasies are morally kosher. If a person regularly imagines having sex with a child or fantasises about raping someone, we might call that into question. Such a fantasy might be thought to express a morally problematic desire. This issue is complicated and depends to some extent on what exactly the person is imagining. In any event, even if we agree that it is morally problematic to have such fantasies, it does not follow that we can or should do anything about it. In particular, it does not follow that we should use the law to police people’s thoughts. We might just think it is morally problematic and people that have such fantasies should critically reflect on their moral significance and, if possible, try to change the content of their fantasies. None of this, however, affects the point that generic or ‘plain-vanilla’, sexual fantasies are thought to be morally permissible.

Here then is the dilemma: What if instead of using their imagination to construct a plain-vanilla sexual fantasy, a person creates a deepfake porn video instead? Does the fact that it has been turned into a material representation make a moral difference? Öhman argues that it should not because it is not obvious how materiality can make a significant moral difference. And yet it seems, when pressed, that we would find something morally problematic about creating the video that we wouldn’t find morally problematic about the fantasy. How can we reconcile these differing moral judgments? That’s the essence of the pervert’s dilemma.

Let me say two things about this dilemma before proceeding. First, Öhman’s claim that turning a fantasy into a material representation is not morally significant seems dubious. Surely there is something morally significant about this? When you create the deepfake, you create a file that is easily shareable and, potentially, weaponisable. We see evidence of this all the time. Deepfake porn videos are both (a) widely shared on the internet and (b) often used to humiliate, degrade and intimidate people (typically women). This shareability and weaponisation is part of what makes them morally problematic and, at least in my own case, is part of motivates me to think they are morally impermissible.

Öhman, to be fair, is fully aware of this point. He sharpens the dilemma by asking us to imagine a case involving a purely private, non-shareable, deepfake. So suppose someone creates a deepfake porn video of someone they know for private use only and there is no possibility of it ever being shared or weaponised or, indeed, of the person whom it depicts ever finding out about it. Would we still find it problematic under those conditions? Öhman argues that we would and this means the dilemma is still relevant.

The problem with this is that it means we are now debating the moral permissibility of a hypothetical technology. I’m not aware of any deepfakes that do not have the potential to be shared nor ones that can be confirmed to be purely private in nature. Furthermore, if such a technology existed, I’m not sure what my moral intuitions about it would be. If you remove the major potential consequential harms of deepfakes, I think I might find them less problematic than I currently do. Debating hypothetical technologies like this can be fun, and might reveal something important about our moral beliefs and practices, but it has limited relevance for the real-world.

The other point I wanted to make about the dilemma is that I think it is badly named. Öhman is drawing inspiration from Morgan Luck’s ‘Gamer’s Dilemma’ (which I have discussed previously). The gamer’s dilemma is about the permissibility of virtual child sexual abuse vis-a-vis virtual murder. But calling his dilemma the ‘pervert’s dilemma’ seems unfortunate to me since the term ‘pervert’ carries negative moral connotations. Ordinarily, when we use the term ‘pervert’, we use it to describe someone with morally problematic sexual attitudes and practices. The whole point of the dilemma is to highlight a tension between something we deem non-problematic (plain-vanilla sexual fantasies) and something we deem problematic (non-consensual deepfake porn). Using the term ‘pervert’ to describe the dilemma confuses things because it suggests that this is only a dilemma for someone with problematic sexual attitudes. A morally neutral title (e.g. the fantasiser’s dilemma) might have been better. That said, I don’t know what the best alternative title would be.

2. Using Levels of Abstraction to Resolve the Dilemma

Set all that to the side. Suppose we accept Öhman’s analysis and focus on the permissibility of non-shareable private deepfakes vis-a-vis sexual fantasies. Is there a way to resolve the dilemma? Öhman argues that there is if we use the idea of levels of abstraction (LoA).

This is an idea taken from the work of Luciano Floridi, which was itself inspired by formal methods in computer science. In essence, the idea is that different objects and entities can be described at different LoAs. Take two human beings: Dave and Emma. You can describe them at the LoA ‘HUMAN BEING’, in which case you focus on attributes they share qua members of the human species. Or you can describe them at the LoA ‘FAMILY’, in which case you focus on their attributes as members of the same family (or not, as the case may be). The former is a ‘higher’ and hence broader LoA because it ignores many properties (variables, observables) that are distinctive of Dave and Emma in their everyday lives. The latter is a ‘lower’ and hence narrower LoA since it adds in additional properties that are more distinctive and unique to Dave and Emma and not just shared across all humans.

Although this example covers the comparison of individual human beings at different LoAs, it should be noted that the method of LoAs applies more generally to all forms of comparison and analysis. The basic gist of the idea is that, for any given question or inquiry, there are more or less appropriate LoAs at which to try to answer that question.

This has implications for morality. It is common in moral philosophy to focus on the permissibility of actions at high levels of abstraction. Philosophers often discuss the permissibility of our actions qua individual moral agents, abstracted away from our social contexts and roles. Consider, for example, the normative principles underlying utilitarianism or Kantianism. These are highly abstract principles, focusing on our duties as agents and/or sentient beings. The problem with this style of analysis is that when we consider our duties at lower LoAs (qua citizens in the same nation or members of the same family), we might find that different duties apply. Indeed, this is a common critique of highly abstract, impartialist moral principles: that they overlook our strong partial duties given the particular social roles and contexts we occupy.

In addition to this, Öhman points out that actions that seem harmless or permissible at high LoAs might be harmful at lower LoAs. For example, hate crimes might be of this form. Considered as an interaction between two abstracted agents, an incident giving rise to hate crime might not be morally harmful or problematic. The victim of a racial epithet, for instance, might brush it off or ignore it or not find it upsetting (the opposite could also be true). But considered at the LoA ‘ETHNICITY’ (or some equivalent) it might be harmful or problematic. The use of the epithet might harm or oppress all members of a given ethnic group even if it does not harm a specific individual. It is also possible that the individual acts, considered at the most abstract LoA, might combine or emerge into something problematic at a lower level.

Öhman thinks that this points the way to a solution to the pervert’s dilemma. In brief, he argues that when considered at a very high LoA, the creation of non-shareable deepfakes might not be morally problematic: the moral agent that created the deepfake (A) might not harm or injure the person represented in the video (B). This makes it, effectively, the equivalent to a private sexual fantasy. But considered at a lower LoA, it is problematic because the creation of the video is made possible by, and feeds into, a system of gender-based inequality and oppression. I’ll quote from Öhman’s article on this point (please note that this quote has been modified to remove some technical terminology that Öhman uses in his article but that I have not explained in this article):

For by abstracting the Deepfake phenomenon into a matter of “A” and “B”, one also subtracts from it the very thing that gives it its ethical significance namely its role in the social system of gender oppression. In one sentence, you cannot take gender out of pornography, and you cannot take society out of gender. As a societal phenomenon, Deepfakes are arguably enabled by a [combination] of male consumers, producers, technology, and misogyny. Moreover, it arguably plays a role in the machinery which systematically reduces women (as a collective identity) to sexual objects, even if none of the individual instances can be held to cause this. So it should be fair to say that the phenomenon is highly gendered (indeed, one need not spend much time on one of the forums or websites devoted to Deepfakes to realise this). While each isolated video may not affect the women it stars as individuals, the phenomenon as such…is, in its current form, inseparable from the systematic degrading of women as a collective identity. 
(Öhman 2020, p 137)


Contrariwise, he claims, sexual fantasies are not like this. While the content of specific sexual fantasies might be inspired by a system of gendered inequality, the general phenomenon of sexual fantasies is not part and parcel of such a system. It is, rather, a common and normal part of human experience.

3. Is this a plausible resolution of the dilemma?

Öhman anticipates and responds to two critiques of this line of reasoning. One of them is that his analysis is just a sophisticated way of saying that the ethical permissibility of non-shareable deepfakes depends a lot on social context and meaning of deepfakes, and hence doesn’t say much of use at all. Öhman is, I think, rightfully dismissive of this critique. His point in using LoAs is that the moral assessment of actions requires some preceding sociology of those actions. This is a fair point and has been made by other philosophers. Purely abstract moral reasoning often doesn’t tell us much.

The other critique is more interesting. It is that his analysis of the impermissibility of deepfakes could apply equally well to other forms of media representing sexual fantasies. He gives the example of a man that draws crude pictures of sexual fantasies instead of using photorealistic deepfakes. Is this morally impermissible? Öhman says that it will depend on the social meaning of such images. He thinks it is unlikely that they have the same meaning as deepfakes:

I believe the answer to this question must be sought in the cultural role of the phenomenon of drawing pornographic images of women one has met. To my knowledge, this is not a common practice used in gender oppression in today’s society, but in a hypothetical society, it certainly could be. 
(Öhman 2020, p 138)


I think this is a bit too quick. When we peel back the sophisticated layers of analysis, it seems to me that Öhman is effectively repeating a very common (radical) feminist critique of all pornographic media, namely: that even if the production of particular pornographic media does not harm anyone involved in its production (and possibly may even benefit them), the media contributes to (or maybe even constitutes) a form of gender-based oppression. In this regard it is telling that Öhman cites the well-known feminist critic of pornography — Gail Dines — to support his analysis of the social meaning deepfakes. So I think his critique could easily apply to other forms of media, possibly including drawings or sketches of sexual fantasies since the very reason why a man might draw such images could be, if we follow the critique, because of some toxic, partriarchal ideology he has imbibed..

What should we make of this? Well, I’ve written a lot about the ethics of pornography on this website before, assessing this common critique from multiple angles. In brief, I would say that the problem with this critique is twofold: (i) it often requires us to overlook, ignore, or explain away examples of pornography that genuinely do not seem morally problematic or harmful to either producers and consumers (and thus it generates a tension between two beliefs about these practices) and (ii) it ignores the fact that the social meaning or significance of pornographic production is not fixed or set in stone. It can change and, indeed, many sex positive producers of ‘feminist pornography’ claim that it may already be changing. Furthermore, when it comes to the social meaning of pornography, it is difficult to completely divorce this from its potential social consequences. If pornography is really harmful to women, and clearly contributes to their ongoing oppression and inequality, then the social meaning of any particular instance of pornographic media is more likely to be negative. If pornography is beneficial or if its effects are more doubtful or unknown, then the picture is less clear.

What does this mean for deepfakes? I think it means two things. First, the analysis that Öhman offers of non-shareable deepfakes is quite similar to, and probably subject to similar critiques as, more general critiques of pornography. Second, it might be harder to defend this analysis if we limit ourselves to the purely hypothetical example of non-shareable deepfakes: whether or not deepfakes are part of a system of gender oppression and inequality depends on how they are shared and weaponised.

This brings us back to the point I made earlier in this article: that debating the merits of a hypothetical form of deepfake porn may have little practical utility.

Tuesday, November 3, 2020

Technology and the End of Reality: Is the infocalypse imminent?

It is now common to hear people fret about the power of technology to distort our perception of reality. With the advent of deepfakes, cheapfakes, fake news, and filter bubbles, it seems that technological forces are aligning to make it harder for us to sort fact from fiction. Some take this fear to an extreme. They worry that advances in technology will bring about the end of a shared sense of reality. No longer will people debate a common core of shared facts and assumptions about the world. Instead, everyone will live inside their own bubbles and dismiss the views of outsiders. We are living, according to these critics, in the shadow of the “infocalypse”.

Will this happen? What are the mechanisms that might bring it about? And does it really matter if it does? In this article, I want to try to explore and tentatively answer some of these questions. Although the claim about the imminent end of reality is common, I’m not sure that it is always well defended. It is easy to point to an emerging technology such as deepfakes and say something provocative about its power to bring about the end of reality; it’s harder to prove that this will actually happen.

I don’t claim to provide definitive proof in what follows, but I do hope I to provide some clarity on how it might happen. In brief, I want to suggest that the mechanisms that could bring about an end to a shared sense of reality are more complex and pervasive than is commonly assumed. This should be a cause of concern though there are some obvious historical parallels to our current situation that should not be overlooked. Furthermore, bringing an end to a shared sense of reality has some attractions, although there is a paradox associated with this.

1. Why is a shared sense of reality important?

Let me start by considering the practical importance of this inquiry. It may seem obvious to you why having a shared sense of reality — by which I mean a belief that you and your peers live in the same world, share some common beliefs and assumptions about that world, and are trying to understand it in good faith — is important but it is always worth asking the obvious question. The answer might be surprising.

The classic philosophical view of humans is that we are reasoning creatures. We have been endowed with senses that allow us to perceive the world around us and an intelligence that allows us to bring some order to these perceptions. We use our intelligence — our ‘reason’ to use the more traditional term — in two distinct ways: to understand (theoretical reason) and to plan and act (practical reason). Furthermore, we don’t just exercise our reason independently and individually. We are social creatures. We reason together, working towards a common understanding of the world and figuring out what we ought to do with the time we spend in it. Having a shared sense of reality is important for all of this.

It is important from the perspective of theoretical reason because it allows us to develop theories and gain insights into the world around us. In order to do this, both individually and collectively, we need that world to be reasonably stable, and capable of being probed and experimented upon from multiple perspectives. To put it another way, we need it to be not entirely contingent on our perception or thoughts. Some degree of dynamism and instability is, of course, inevitable, but as long as it is orderly and pattern-like we can still hope to bring it within the scope of theoretical reason. Some element of perceptual distortion is inevitable too. Our minds are not always good at gaining an accurate impression of the world around us, but through effort we can collectively limit the distortions and make reality sensible. None of this works if we all assume we are working with different realities, or if the world around us is, in fact, chaotic, disorderly or highly mind-dependent.

A shared sense of reality is also important from the perspective of practical reason. In order to make plans and implement them, we need a world that is reasonably predictable and stable. If the world is not stable and predictable, plans go out the window. What’s more, this can also have an impact on the moral dimension of practical reason. We need a shared sense of reality in order to develop and follow moral rules. There are several reasons for this. First, most moral norms assume certain stable facts about the world we live in. Consequentialist and utilitarian moral norms, for example, assume that certain actions have predictable effects on human (or sentient) suffering and pleasure. Kantian moral norms assume that we live in a world inhabited by other agents that experience and act in the world in a similar manner to ourselves. Without assuming these external facts, most moral theories dissipate into nihilism. Similarly, when we wish to speak truth to power and address moral injustices or outrages, we rely on the ability to get others to ‘see’ some aspect of the moral reality that they were previously ignoring or overlooking. Ending slavery, for example, relied on getting others to see that we shared a common humanity and hence suffered and experienced the world in a similar way. Seeing this shared reality was key. Again, some dynamism and probabilistic uncertainty is inevitable, and does not completely scupper our moral projects. We just need a reasonable degree of stability and order. 

More reasons could be adduced but for now this should suffice for now. It can be added, as well, that I don’t think any of the claims just made rests on particularly controversial philosophical foundations. A lot of this is because I’m focused on our shared sense of reality and not on the more controversial claim that there is, in fact, a single shared reality that we are capable of knowing. Maybe there is; maybe there isn’t. Perhaps we can only ever access imperfect representations — shadows on the cave wall — of reality. Perhaps the Buddhists are right and reality is ultimately formless and chaotic. I don’t think this affects the claims I just made. It would still be the case that having a shared sense of reality would be valuable for theoretical and practical reasons.

2. Is Technology Undermining a Shared Sense of Reality?

Assume what I have just argued is correct. The next question is whether technology really is undermining the shared sense of reality. The answer to this, I think, is that it certainly has the power to do so, but it doesn't necessarily do so, and if it does it does so in tandem with other causal factors, in particular with human psychology and social institutions. In other words, it is the combination of these three forces — technology, psychology, and society — that poses the threat to the shared sense of reality not technology alone. Allow me to elaborate.

Consider the technological mechanisms that might undermine a shared sense of reality. Three seem particularly important:

The algorithmic curation of information: It’s a banal observation but it bears repeating: we live in the information age. Never before in human history has so much information been collected, stored, organised and broadcast to human beings. So much so, in fact, that it is impossible for us to make sense of it without considerable technological assistance. We now rely, daily, on algorithmic platforms (Google, Facebook, Netflix, Amazon, Twitter etc) to curate and make sense of this information. These platforms often personalise and adapt the information that they present to us and our peers, creating filter bubbles and echo chambers to cater to our informational preferences. The end result is that we increasingly live inside informational hubs that reflect our preferred perception of reality and not necessarily the general and shared perception of reality.


Deepfakes and other easy forms of media manipulation: Media has always been manipulated. From the first tablets and scrolls to the printing press and beyond, people have always sought to create fake or forged documents, photos and movies. What’s happened more recently is that, with the rise of Deepfakes and other, easy-to-use forms of media manipulation, the power to create hyperrealistic fake media has become more widely distributed. This degrades our informational commons. As the philosopher Regina Rini has argued, the rise of deepfakes (and their cousins ‘cheapfakes’) removes an honesty check we have on the truth or falsity of testimony: we cannot use video to speak truth to power if video is easily faked. Similarly, and more generally, Don Fallis argues that the rise of deepfakes reduces the amount of information carried by audiovisual signals by increasing the likelihood of their being false positives. The end result is a world in which it is easy to both believe in the media you prefer to believe in and dismiss the media you don’t wish to believe in, on the grounds that it is likely to be fake.


The emergence of immersive and realistic forms of virtual and augmented reality: This is, in a sense, the apotheosis of the filter bubble. Instead of just relying on highly curated, possibly faked stories and audiovisual records, we now also have the power to immerse ourselves in absorbing (if not yet hyperrealistic) virtual worlds and to overlay computer generated images and other content onto our perception of the real world. These technologies can often trick the senses into thinking (if only for a moment) that the virtually constructed reality is the same as the real thing, and they give people an escape valve from the drudgery of the real world. The end result is that people can escape into their own preferred worlds when they desire, and do not need experience the same reality as others even when they don’t.


Many discussions of the 'infocalypse' or the end of a shared sense of reality tend to focus on one of these technological developments to the exclusion of the others but it is the combination of them that is the problem. Furthermore, these technological forces are complemented by the other mechanisms that I will now discuss.

Consider the psychological mechanisms at play. As Henrik Skaug Saetra points out in his article ‘The Tyranny of Perceived Opinion’, the algorithmic curation of information plays off certain psychological biases that most people share. In particular, it plays off the related phenomena of selective exposure and confirmation bias. These are well-evidenced psychological biases. The former describes people’s tendency to only notice or seek out information that supports their pre-existing values and perceptions of the world; the latter describes the tendency to interpret any dissonant or incongruous information in a manner that confirms pre-existing biases. It is these psychological tendencies that push both users and platform creators into filter bubbles and echo chambers. This is connected to how people use emotions to evaluate and express their perception of the world. As Steffen Steinert argues in relation to the phenomenon of emotional contagion on online platforms, we all use emotions to attach values to our perceptions of the world. We also share these emotions with others and feed off other people’s emotional responses. If you are afraid of something, and clearly express that fear, my own fear response is more likely to be triggered. Emotions are important. Without them we couldn’t select and sort information in a useful way, but emotions can also distort our perception of reality. The problem with technology, particularly social media, is that it promotes ‘hot’ emotional responses to the world. Expressions of outrage and anger are more likely to be seen by more people. When people are angry and outraged they are more likely to adopt a conservative, closed-minded view: suspicious of others and protective of themselves. This further promotes filter bubbles and echo chambers.

Finally, consider the institutional mechanisms at play. By the use of the term ‘institutional mechanisms’ I mean to refer, primarily, to political and economic institutions. The economic actors that create many of the informational tools we currently rely upon — and that are undermining our sense of a shared reality — rely on a particular business model that encourages them to both appeal to and promote a polarised and fragmented view of the world. Multiple news organisations now exist to ’niche’ themselves to particular segments of popular opinion. The same is, increasingly, true of political institutions too. Political parties niche themselves by appealing to particular fragments of popular opinion. Ezra Klein has discussed the net effects of both institutional forces in his book Why We’re Polarized. Though this book is US-centric in terms of its focus some of the key insights apply more generally. At the heart of Klein’s argument is the claim that increased political polarisation is being driven by a feedback cycle: the population is becoming more fragmented (for a variety of reasons, including those having to do with the fragmented and personalised media landscape), and political institutions are responding to this by appealing to the fragments. This has a snowball effect: when one party is in power, they appeal to their fragment of the population, and this amplifies the sense of ‘identity threat’ among rival fragments who respond in kind with increased attachment to their preferred values. The process cycles on, resulting in an increasingly polarised political climate. The polarised communities see outsiders as belonging to a different world and a major threat to their own existence. It is worth noting, as well, that different factions can nefariously manipulate the polarised informational environment to suit their own. Many governments, most notably Russia, do this to foreign countries in order to support their own power and influence.

These three mechanisms feed off one another and have two main effects:

E1: It is now much easier for individuals to slide into their own preferred construction or perception of reality (and not encounter or be disrupted by dissonant information).


E1: It is now much easier for people to dismiss or doubt the reality presented to them by others (on the grounds that it might be faked or distorted in some way)


Both of these effects undermine a shared sense of reality.

3. Is It Really So Bad?

The preceding paints a bleak picture. We might wonder whether things are really so bad. I close with three observations.

First, as noted above, media manipulation is nothing new. Humans have always manipulated media to various ends. Furthermore, the psychology of selective exposure and the problems of political polarisation are hardly unprecedented in human history. They have always been with us. What is usually earmarked as being different this time round is the volume of information with which we must contend, the relative ease with which it can be manipulated, and the realistic nature of the end product. But part of me wonders whether things really are so different from what they were once like. One of my hobbies is biblical history, particularly the history of early Christianity (I’ve written about aspects of this before on this blog). If you approach that era from a secular perspective, one thing that often strikes you about it is how epistemically chaotic it seems to have been. Different religious sects with different worldviews were commonplace. People within those sects listened to stories told to them by friends of friends of friends that might have witnessed the events that originated the sect. They often adopted life-changing worldviews on the basis of this imperfect testimony. What’s more, when these stories were written down they were often presented in forged documents (in the sense that the author’s pretended to be people they obviously were not) and occasionally mistranscribed or misedited, distorting some crucial element or lesson. Lots of rival, distorted and dubious stories fought with one another for attention. People believed on thin evidential foundations. I’m sure much of the ancient world was similarly epistemically chaotic. What might be happening now is that technology is just returning us to the epistemic chaos from which we briefly escaped. This doesn’t make it a good thing: a step back is a step back. But it may not make it a radically new thing either.

Second, as Henrik Skaug Saetra pointed out to me, the argument I have presented doesn't necessarily undermine a shared sense of reality. It could well be the case that technology causes greater fragmentation of the human population into different bubbles. But unless those bubbles are completely individualised, it's likely that we will share a sense of reality with the other occupants of our respective bubbles. For example, people living in, say, the left-leaning liberal bubble on Twitter are likely to share a lot of common beliefs about the nature of the world they inhabit. Indeed, it may even be the case that they share more beliefs than would previously have been the case since they primarily associate with one another. The problem arises more at the inter-group level. The techno-social infrastructure is now such that there is greater disparity and dissonance across the bubbles than within them. All of this sounds right to me and may prevent the slide towards complete epistemic chaos, but I nevertheless think it still undermines a shared sense of reality since there may be many more fragmented bubbles than before. This is because, instead of having to figure out some common modus vivendi with our neighbours and fellow citizens, technology affords us the opportunity to retreat into a world of our own making.

Third, and finally, there is one potential argument for thinking that the end of shared sense of reality is not that bad One of the arguments I experimented with in my book Automation and Utopia was Robert Nozick’s defence of the idea of a meta-utopia. In short, Nozick argued that people will never agree on what the best world or best life is. We should stop trying to get them to do so. Instead, we should create a meta-utopia which is a world in which people are allowed to create and join mini-worlds that suit their own values and preferences. In Automation and Utopia, I argued that one of the advantages of technology — including technologies like AR and VR — is that it may allow us to create a functional meta-utopia: people can use technological tools to build worlds that match their preferences. However, there is, as I pointed out, a bit of a paradox in this. The meta-utopia is only desirable if people respect the boundaries between all the mini-worlds — i.e. they don’t seek to undermine or ‘colonise’ another person’s preferred world. How can we guarantee this? The only conceivable way is if we agree on some general constitutional order that polices the boundaries between all the mini-worlds. So, ironically, we all have to share enough of the same sense of reality to create that constitutional order before we can happily escape into our own realities.