key: cord-0060130-hocw4x0q
authors: Myketiak, Chrystie
title: Automating Desire
date: 2020-10-10
journal: Online Sex Talk and the Social World
DOI: 10.1007/978-3-030-53579-7_6
sha: 134886ba649c9e9e192c657b188af37bdb5d11f3
doc_id: 60130
cord_uid: hocw4x0q

This chapter examines automated commands that Walford participants have created to simulate sexual practices. Although the commands are overtly erotic, the analysis demonstrates that rather than serving such purposes these commands serve multiple in-group ludic functions. Their function as mimesis may be most outwardly evident, but it is their use as games of chance that governs how they are deployed and understood by members of the community. Although these commands may be interpreted as games by participants, it is reasoned that they are situated within sociosexual discourses where sexuality is quantifiable, consent is dependent upon opting out, and sexual agency is dependent upon only one participant in a dyadic encounter activating their sexual agency over the encounter, however mutual the scenario may appear.

Web-based chatbots are programmable bits of code that will produce text or talk responses dependent upon the information that they receive. Chatbots serve multiple prosaic functions in a neoliberal capitalist and increasingly automated society. For example, chatbots can be used by companies to engage in 'live chats' with customers in the first line of customer support, assisting customers in choosing and comparing products, reporting missing products, or updating information. Chatbots may also operate as virtual assistants, with Apple's Siri and Amazon's Alexa being two examples of well-known commercialised virtual assistants that are programmed to interpret and respond to human language. Siri and Alexa have the capacity to answer questions, control home automation devices, and manage organisational tasks such as emails and calendars. While they are examples of sophisticated bots in that they both interpret speech and respond via (synthesised) voice, there is a wide range of linguistic functionality in bots. Some, like those in Walford, have predefined linguistic functions that produce lines of text when activated, while others are highly sensitive, with developers incorporating artificial intelligence technologies so that the bots respond to the information they receive. In theory these highly sensitive bots can accumulate intelligence or at least filter and categorise the information they collect. These bots can be interpreted as developing communicative competence based on their experiences of information. Just as someone who uses a music listening application can expect their recommendations to become more personalised and tailored to their existing listening habits over time, the idea is that after extended use or deployment, a bot that is written to adapt to external information will become more attuned to the linguistic landscape surrounding it. For example, Amazon used artificial intelligence to sort through job applications and over time their technology made the choice to remove applications that included the word 'woman'. The software was designed to recognise that applications containing this word were less likely to be selected by humans and, not understanding bias behind it, began to filter the applications in the same way. When Amazon executives discovered that this was occurring, they disabled the technology and vowed to tackle the inherent selection bias that it reflected and exposed. The point is that technology is never neutral: it is created by people, is used by people, and exists within a broader political and sociocultural realm. Technological innovations reflect the biases that exist among humans because the ways in which engineers and designers-like everyone else-understand the world is through a specific lens, and artificially intelligent artefacts will develop the same biases that exist in the wider ecosystem because they make sense of the information that they receive as input to respond accordingly. Windle (2014) suggests that chatbot-human conversations recall a form of misanthropy that is often found in binaries, such as that of master/slave or dominant/subservient. Yet, one important aspect to consider in the interactive process of human-chatbot engagement is how this misanthropy may provide an opportunity for asking questions about how inequality and discrimination are programmed into the social landscape, how they recreate them, and what steps can be taken to challenge or eliminate their propensity for reproducing inequality and discrimination. Concerns about chatbots becoming echo chambers for oppressive discourses are not abstract or theoretical; chatbots can and have developed racist, sexist, homophobic, and anti-immigration stances. This can happen because bias has been included in how the bots have been programmed or through a combination of their programming and their interactional environment, which has led the bot to view this is normalised. In either case they are repeating cultural scripts, phrases, and stances that they have received from humans and interpreted as input. One example is Tay, a chatbot created by Microsoft, who was on Twitter for fewer than twenty-four hours before becoming racist, sexist, and anti-Semitic. As Schlesinger et al. (2018) explain, this could have been avoided if developers had deployed a blacklist or wordfilter, which would have excluded filter-specific words and sub-strings from Tay's vocabulary. On the surface, a blacklist seems like an easy and neutral way to create a bot that is-by design-anti-oppressive. However, blacklists can become an ethical quagmire. Some words and strings may be included without much debate or controversy, (e.g., the n-word, faggot) but others are more contentious, especially if the code is programmed for strings rather than words, (e.g., fag, homo). For example, the string homo excludes the original Latin term for human beings and would have the unintended consequence of removing discussions about homo sapiens and homogenisation and fag may also refer to a cigarette. Although a designer may have good intentions, that doesn't mean that the design is not problematic or that issues of voice and representation have been addressed in the product that has been developed. One workaround that has been written into some blacklist code is excluding the first syllable of words from filtering. However, this too may have unintended effects. In 2020, Dominic Cummings, chief adviser to the UK Prime Minister, caused outrage to many in the United Kingdom and across the world as a result of his movements during the Covid-19 pandemic. Twitter users found that any hashtags with his complete surname were not trending due to Twitter's anti-porn filters. Misspellings of his name in hashtags were adopted in order to have the topic trend; however, users discovered that the antiporn blacklist excluded the first syllable of words to differentiate words like cumbersome and Cumbria from others such as cumming. This meant that #cumgate was able to trend while #dominiccummings could not.

The infrastructure of Walford provides participants with opportunities to create new communication types and settings within the multi-user domain (MUD) and also the option to implement new automated commands which, when activated, automatically enact an action or dialogue. Users have been remarkably creative in the automated commands that they have produced; however, the bots are representative of the technological interface of MUDs. Participants created automated commands to provide themselves with announcements for when they enter and leave rooms. Other automated commands announce when users connect and log out of Walford. Both of these types of commands were discussed in Chap. 4. Users can change these as often as they wish and use them as forms of personal expressions and branding strategies. In an active community like Walford where there are approximately 1500 regular users, a quote upon entry or exit can be a form of 'presentation of self ' (Goffman 1959) that distinguishes one user from others in a text-based environment. For example, one user has a series of Confucius-inspired quotes appear when they enter or leave a room, and these short quotes become embedded in their indexicality in much the same way that another user's exit statement that references sex workers, gambling, and amusement parks is encoded with how he wishes to be seen by other Walford users.

While discussions of sex are formally discouraged in the rules of Walford and participants are advised that exchanging 'pornography' will result in expulsion from the community, two of the most popular automated commands are shag and kiss. Although it could be argued that these commands have slipped through the community regulations or that administrators have chosen to overlook them in order to acknowledge the value they bring to the community's social cohesion, these commands simulate erotic practices but are treated by participants as neither sexual nor erotic. As a result, they are not viewed as problematic or counter to the community norms or standards.

The automated kiss and sex commands generate erotic scenarios between the participant who initiates the command and the user who the initiator implicates, with both the initiator and recipient needing to be co-located within the same Walford room. These commands are written so that mutual engagement is implicit in them even though it is one participant who activates the command towards another user without their knowledge or (explicit) consent. Unless a participant takes the step of blocking the commands (i.e., opting out), others are able to implicate them. If a user does not block the commands, they have appeared to have 'opted in' to the possibility of another user activating it towards them at any time. Differences between opting in and opting out were discussed in Chap. 3, and this distinction is realised in these overtly erotic commands here.

Despite this complex area of consent, the command is written so that both participants are equally engaged in the process although the user who has activated the command is the agent. This kiss has two generated components: the type, which is generated from a closed set of types, and the length of time that it leaves the participants breathless. The shag command, similar to the kiss command, produces a sexual position that two participants share and its duration. There are five points regarding these commands to draw out now. First, these commands have been programmed to simulate actions between two agents and have not designed to incorporate either solo or group actions. In this regard, the activities are replications or copies of normative dyadic sexual scenarios and can also be interpreted as falling within the 'good, charmed circle' of Rubin's (1998) sex hierarchy. Second, the commands rely upon an active participant who initiates the command and a passive recipient who must not have blocked the command. The passive participant's consent to be involved in a kiss or shag is implied through the failure of blocking the command. Third, although these commands describe a type of kiss or a sexual position, they are generally used as play in the ludic sense, rather than as flirtatious or erotic enactments. Fourth, users have developed conventions for using these commands that are contextualised in the lines both preceding and succeeding the automated command lines. One of these is a pattern of adjacency with convention dictating that the receiver of the command reciprocates the command towards the previously active user. Fifth, the commands interpreted as play, rather than as erotic gestures and they do not lead to erotic overtures.

The kiss command enables an automated string of two lines of text surrounding a kiss between two Walford participants who are co-located in the same room. In order to execute the command, the user types the word "kiss" followed by the name of the user that they wish to initiate this activity with. Example 6.1 refers to the content that is generated by this command (two lines), and emphasis is added to stress the content that is variable.

1 20703 pulls 17212 close for a sweet sexy kiss 2 20703 kisses 17212 until they are breathless for 12 seconds 3 17212 pulls 20703 close for a passionate kiss 4 17212 kisses 20703 until they are breathless for 19 seconds 5 17212 I beat you

In each line the aspects of the command that vary are the adjectives used to modify the kiss, giving it dimensionality and constructing a fantasy for it, as seen in Lines 1 and 3 ('sweet sexy' and 'passionate') as well as time specifications, as seen in Lines 2 and 4 ('12' and '19'), though 'seconds' rather than minutes, days, or months are used to mark time. 20703 directs the kiss command towards 17212 and the command randomly generates a type of kiss from a constrained and short list of possibilities; 'sweet sexy' in this case. The line that immediately follows marks the length of time that the participants are 'left breathless' following the kiss; '12 seconds' here. Although it is clear who initiated the command and there was no consent or opting in before the command was activated, the kiss' code is structured to imply consent and mutual engagement through the shared third-person pronoun 'they'. In addition, one participant is left breathless no longer than the other, which also seems to imply that in this scenario the users imagine that there is consent, shared engagement, and equitable levels of pleasure. Although there are a number of different styles or types of kisses that can be produced from generating the script, neither refusal nor rebuttal of any kiss has been programmed as a possible outcome. When these features of automated kisses are combined together, the command not only seems playful but the element of fantasy is heightened. There are no possible negative outcomes built into the code, and while users have the ability to create these possibilities and add them to the command, this has not happened. The command is distinctly separate from the 'real' where social actors may respond with refusals and may also experience the same kiss in different ways. Huizinga (1992: 8) argues that "play is not part of 'ordinary' or 'real' life. It is rather a stepping out of 'real' life into a temporary sphere of activity with a disposition all of its own". Play then can be understood as essentially indulgent, gratifying, or pleasurable because it removes the player from the confines of 'real' and into an 'imaginary' sphere or scenario. Caillois' (2001: 5-6 ) point about play is similar to Huizinga's when stating, "play is an occasion of pure waste: waste of time, energy, ingenuity, skill, and often money […] There is also no doubt that play must be defined as a free and voluntary activity, a source of joy and amusement". The automated kiss command can be understood within both of these understandings of play. When a participant activates the command, they are engaging in an activity that is indulgent and "a source of joy and amusement". The kiss script is a Walford convention that participants enjoy, and one that they engage in freely and voluntarily. It involves stepping into an imaginary realm that is more than the fantasy scenario of the kiss because it produces a positive affective state for themselves and the people they engage in this play with.

Caillois (2001) breaks games into four types: agon (competition), alea (chance), mimicry (simulation), and ilinx (vertigo). He gives the following examples: "One plays football, billiards, or chess (agon); roulette or a lottery (alea); pirate, Nero or Hamlet (mimicry); or one produces in oneself, by a rapid whirling or falling movement, a state of dizziness and disorder (ilinx)" (Caillois 2001: 12) . Using this typology, the automated kiss command can be seen as a ludic activity, more specifically one of mimesis or mimicry, such as that which is associated with the parlour game Charades. In Charades, players act out words or terms whilst other players attempt to guess the iconical representation. The kiss command works as a form of mimesis that involves participants simulating the act of kissing through text and creating multimodality for it through the combined use of adjectives, time duration, and affective state. However, the command is not only mimesis: it simultaneously acts as an alea or a game of chance. The type of kiss and the length of time for which the participants are left breathless are randomly generated, and the analysis shows that while the mimesis aspect of the game is not typically commented upon, participants regularly comment upon the quantifiable aspects of the kiss.

How participants react to the alea aspect of the game entwines it with a sense of competitiveness, though it lacks the strategic competitiveness inherent in billiards and chess. Participants tend to engage in dialogic resonance with respect to the command, initiating it reciprocally after someone has initiated a kiss with them. The competitive features are highlighted in subsequent turns when the participants designate a winner and loser of the game. Returning to Example 6.1, in Line 5, after the cointerlocutors each takes a turn at initiating the command, 17212 states, "I beat you", making the point explicit. The 'passionate kiss' that they initiated in Line 3 led the participants to be breathless for seven seconds longer than the 'sweet sexy kiss' initiated by 20703 in Line 1. Thus, the aleatoric feature of the command fosters the competition between the users that rests on time duration rather than a qualitative function that may require participants to disclose (or believe that they are disclosing) something about their personal desires, erotic tastes, or sexual preferences. In this sense, the convention for reciprocity as seen in adjacency pair usage of the command is related to the competitive value that the participants associate with the kiss command. The comments that result from comparing the duration of breathlessness of the kiss function as a way to respond to the action without needing to treat the kiss as if it were either a kiss or a simulation of one. This point becomes clearer in Example 6.2.

1 3293 pulls 14863 close for a lingering kiss 2 3293 kisses 14863 until they are both breathless for 10 seconds 3 3293 That's pants 4 3293 pulls 14863 close for a whisper like kiss 5 3293 kisses 14863 until they are both breathless for 17 seconds 6 14863 pulls 3293 close for a lingering kiss 7 14863 kisses 3293 until they are both breathless for 17 seconds 8 3293 Well I am out of touch today then 9 3293 I can't even snog 10 3293 Pouts

Here the alea aspect of the command becomes more apparent. 3293 rejects their first attempt with the statement "That's pants", with 'pants' being British slang for "not very good". 3293 then enables the command again as if it were a lottery, with the hope to improve upon the ten-second breathless performance. The user is more successful with their 'whisper like kiss' which leaves the participants breathless for seventeen seconds rather than the ten seconds that the 'lingering kiss' in Line 1 achieved. The co-participant of this kiss then reciprocates with a breathless score that matches 3293. Yet, there is also an agon aspect to the interaction, even though it lacks strategy. 3293 remains concerned with their selfperceived poor performance time and expresses this in three consecutive turns (Lines 8, 9, and 10). The first turn is bookmarked by discourse markers, with both 'well' and 'then' serving those functions as neither is necessary, though they add emphasis. 3293's use of "out of touch" to refer to a poor performance is clarified in the second turn when they state, "I can't even snog", using snogging as a barometer for functionality. Finally, the third turn conveys an expression (pouts), which itself is a representation of an affective response, used to bookmark the speaker's belief that this is a competition in which they have performed poorly.

In both Examples 6.1 and 6.2 the users focus on ludic issues, specifically self-performance. The kisses are treated as quantifiable objects, particularly by those who enable them, even if a user has just received and then enabled the command. Furthermore, the users assume that the game is 'won' by achieving a result, purely by chance, of a length of time of the breathlessness rendered. In these two examples the lengths of ten and twelve seconds were deemed to be poor and in the latter the duration of seventeen seconds is also defined as inadequate. Crucially, users do not comment on the type of kiss or attach meaning or symbolism to this in terms of game or other play. Whether a kiss is generated as 'sweet sexy', 'deep passionate', 'lingering' 'passionate', 'teasing', or 'whisper like' is of little significance to Walford participants, who rarely if ever remark upon the kiss type generated. Furthermore, the kiss command is not used as part of or as a precursor to other types of non-automated sexual conversations. Yet, participants do use this command before enabling the shag command. The sequential pattern that users enable the commands confirms normative rituals around erotic practice that are engrained in notions of romantic love as discussed in Jenkins (2017) and in Chap. 2. Despite the rituals surrounding how participants deploy these commands, the forms and features of the kiss command indicate that it is used as a playful and competitive game activity in Walford, but not as a form of erotic activity. In contrast with Examples 6.1 and 6.2, the emphasis is on competition against others in Example 6.3.

1 44639 pulls 1986 close for a sweet sexy kiss 2 44639 kisses 1986 until they are both breathless for 24 seconds 3 44639 Cuz you make me want to do that [a short playful conversation follows] 4 1986 pulls 44639 close for a teasing kiss 5 1986 kisses 44639 until they are both breathless for 13 seconds 6 44639 pulls 1986 close for a deep passionate kiss 7 44639 kisses 1986 until they are both breathless for 2 seconds 8 1986 2? 9 1986 prods 44639

In addition to the playful competition that is evident in the speakers' communication in Examples 6.2 and 6.3, here an emotive element is also present though not quantifiable. In Example 6.2, 3263 sulks in response to a self-perceived poor turn at the automated command, whilst here in Example 6.3 the passive participant jests about the duration of breathlessness (Line 8) through the question "2?" and takes another turn to "prod" or poke their co-interactant in the command. The comment in Line 3 stands out from other turns that follow the kiss command in the dataset. When 44639 states, "cuz you make me want to do that", immediately after having enabled the command, it is possible to see that while the command is not used for explicitly erotic or sexualised purposes within Walford, it can be strategically deployed as a transition to flirtatious dialogue. In this example, the kiss command operates as a tool that 44639 uses in order to transition to an utterance that conveys sexual desire, operating in this function because the automated command does not convey that affective state. In other instances, participants use the command for its creative kinaesthetic purposes. Here 14863 uses the command as part of a goodbye (Lines 3 and 4), with the command functioning in much the way that a kinaesthetic emoji would (e.g., kiss face emoji, dancing lady). As with kinaesthetic emojis, the use in this context is to provide a representation of an action. The distinction between the perfunctory usage of the command and a user-generated 'kiss' is highlighted in the following turn. 17933 completes the adjacency pair by kissing 14863 back, but rather than it being a kiss modified with one adjective as is standard with the command, she uses a collection of twenty-four modifiers (possibly twenty-five, if a comma is missing between 'OhOhOhGawdOhGawd' and 'Soul Stealing') with emphasis put on each word through capitalisation. The phrases that she uses range from standard adjectives (e.g., deep, hard, long, hot) that could be used to describe other sexual acts and body parts as well as those that express her individual linguistic erotic style as quick-witted and clever (e.g., Dream Making Close to XRated, Better Than Heaven, Yet Hotter Than Hell).

The automated shag command, similar to the kiss command, is an automated command developed by Walford participants that is specific to the Walford ecosystem and, despite the overt erotic meaning, conveys little if any erotic desire between participants. It is the disconnect between the erotic performance of the command and how it is read as ludic that makes it serve an in-group function to participants. The verb and adjective laden narrative compresses the two lines of the kiss command into a single line (Line 3) which contains both the position of the sex act and its duration. Similar to the previous command, the shag command communicates two pieces of descriptive information that are randomly generated from a closed set of possibilities (though users are able to add to the existing options): the position or setting in which the participants 'shag senseless' and the duration of the act. At the time that the research was carried out the options included: a position dedicated to the user initiating the command (e.g., in the 32282 position), missionary, doggy style, acrobatically, and like Chapter 4 from the Kama Sutra. Most of the options adhere to a formulaic structure that is prefaced by "in the" which is followed by the randomly generated adjective and the word "position": in the front seats of the Mustang, in the back seats of the Mustang, in the backyard, in the shower, on the edge of the bed, and against the wall. Options that foreground location tend to use the definite article in their description (e.g., the Mustang, the backyard), which creates erotic imagery that grounds and centres the command.

The first line of the command makes it clear that the participant who activates the command 'undresses' their co-interactant and 'caresses' them. The use of "all the sensitive areas" relies on pragmatic inference. The reader must need to have a conceptual understanding that codifies 'sensitive areas' as erotic. The use of the article (the) obscures the person reference but in doing so requires the targeted user to infer that 'the sensitive areas' are on their body rather than an immaterial object. In the following line, again, it is the user who activates the command who "plays" with their co-interactant, although "both [interactants] get more and more aroused". The same pattern is present in Line 3 with the user who activates the command presented as having sexual agency in the scene as this person 'shags' the other. In Line 4, the interactants have equal weighting though the verb here is "collapse", which counters the mutually involved actions in the previous turns.

The structure of the first three lines is that the first word of the turn is a verb and it is the communication structure of the command that leads the user who activates it to have agency here. However, the formula changes for the coda of the narrative in Line 4. Instead of a verb, the first word here is the conjunction "and" which allows both users to share the verb and its agency. This structure could have been used throughout. Instead, there is an unequal distribution of agency and action in the command: one user "undresses" another (what about the other user's clothes?), one user "plays" with the other (both users are "more and more aroused" by this non-reciprocal activity) and despite this the narrative suggests that there is no issue. Reciprocity exists not in the structure of the command's output lines but in how the participants use the command.

1 11116 undresses 5486 slowly and seductively, caressing all the sensitive areas 2 11116 plays with 5486 until they both get more and more aroused 3 11116 shags 5486 senseless like no one has ever seen for 275 minutes 4 11116 and 5486 both collapse in a completely satisfied state 5 11116 Is that all you wanted ;) 6 5486 Pretty much, but my ass is far more sore than it should be 7 11116 Sorry 8 5486 undresses 11116 slowly and seductively, caressing all the sensitive areas 9 5486 plays with 11116 until they both get more and more aroused 10 5486 shags 11116 senseless in the missionary position for 61 minutes 11 5486 and 11116 both collapse in a completely satisfied state. In this exchange the participants enable the shag command four times forming two complete adjacency pairs: 11116, 5486, 11116, 5486 . In addition to the reciprocal use of the command, the participants joke with each other about their use of the commands. This is established in Line 5 after 11116 first shags 5486 and then asks, "is that all you wanted ;)". The humour here is in the idea that activating the automatic command serves the same purpose as a hook-up. While the use of the commands and the resulting comments in Lines 6, 12, and 25 could be read as erotic, the analysis doesn't lend itself to that interpretation. In Lines 14-16, for example, the participants work to develop intersubjective alignment based on heterosexual homosocial masculinity and a resultant theorisation of gay sex. The turns in Lines 12 and 25 rely on heterosexual pornographic tropes in comments about the duration of the automated sex, and these references call upon participants' shared knowledge and understanding.

The use of the shag command here, particularly in the way that participants use it repeatedly in reciprocal turn-taking, is evidence that the command is not used in a way that promotes sexting. This is further supported by a lack of evidence that I found in the chatlogs that the command is used as a precursor to cybersex, flirting, erotic webcamming, or other linguistic sexual practices. Both the shag and kiss commands appear on the decontextualised lexical level as erotic though the participants use and recognise them as playful. At times this playful use of the commands is competitive, echoing the game theory work of Huizinga (1992) , yet at other times the competitive elements are suppressed for a more jovial discussion of the possibilities that are produced by the commands.

While the automated kiss and shag commands are rarely if ever used by Walford participants as precursors for erotic conversations, participants may activate the kiss command prior to using the shag command. Although the use of these overly sexual, seemingly erotic simulations of sexual activities are not read as sexual within Walford, when participants use them in these ways or when they attempt to inscribe corporeal bodies onto them as seen in Example 6.7, participants understand the commands within larger sociosexual discourses that regard them as erotic practices.

1 23639 pulls 5486 close for a sweet sexy kiss 2 23639 kisses 5486 until they are both breathless for 579124448 millenniums 3 23639 undresses 5486 slowly and seductively, caressing all the sensitive areas 4 23639 plays with 5486 as they both get more and more aroused 5 23639 shags 5486 senseless on the front seats of the Mustang for 78 minutes 6 23639 and 5486 both collapse in a completely satisfied state 7 23639 Damn idling ;) 8 23639 pulls 5486 close for a deep passionate kiss 9 23639 kisses 5486 until they are both breathless for 496425511 millenniums 10 5486 pulls 23639 close for a whisper like kiss 11 5486 kisses 23639 until they are both breathless for 25 seconds 12 5486 undresses 23639 slowly and seductively, caressing all the sensitive areas 13 5486 shags 23639 acrobatically for 27 minutes 14 5486 and 23639 both collapse in a completely satisfied state

Here participants engage in the reciprocal use of the commands that characterise their deployment within the Walford context; however, there is an ordering pattern to the use that reflects normative temporal patterns of desire and sexual practice. This same patterning (kiss-shag) can be seen throughout the Walford dataset when both sets of commands are activated in a turn-taking pattern. Horowitz and Spicer (2013) summarise more than ten years of research investigating what types of sexual practices people consider as "having sex" (see also Chap. 1). Across the various studies they sample, vaginal intercourse is most likely to fit within perceptions of what defines 'sex' while kissing is the sexual behaviour least likely to coalesce with the participants' mental categories of 'sex'. This research provides empirical evidence that relates to everyday metaphors about sexual activity. The baseball metaphor that connects sexual behaviour to 'bases' is one example where sexual behaviours are both separated and built upon a trajectory in which "having sex" functions as the ultimate act. Horowitz and Spicer's (2013) work, which asks individuals to define sex, and the baseball metaphor show how there is a general and unmarked form of sex that most people call to mind as well. What is also evident though is that this unmarked form that serves as the generic and default exists within a semantic field that contains a series of other related practices, with each of these requiring explicit naming and clarification from the unmarked form. These differentiations and gradients of sexual activity are manifested in the use of these commands, which are used for ludic rather than erotic purposes. This shows that even in play heteronormative sociosexual discourses and narratives are pervasive. Kissing and sex are treated as distinct by the participants, and additionally, there is a spatiotemporal ordering to the use of the commands that reflects participants' perceptions of sexual activity. This patterning is evident in how participants choose to deploy these commands even when the commands themselves have been creatively augmented, as seen in Lines 2 and 9 when 23639 kisses the co-participant until they are both breathless for hundreds of millions of millenniums. While the commands are used outside of sexualised contexts and function within group-specific norms and expectations, the larger sociosexual discourses about sex and sexuality that ground the commands remain.

This chapter situates Walford's automated sexual commands within broader debates around chatbots and artificial intelligence and focuses on how participants use automated commands that simulate the sexual acts of kissing and sex. These commands are overtly erotic in their structure and design, but there is a narrow and specific form of sexuality expressed through them. Issues of consent, and the difference between enthusiastically opting in and failing to opt out, can be seen in how the commands function. The performance of heteronormative desire is structured into the narratives that these commands create, and this is evidenced in their dyadic nature as well as in how participants spatiotemporally order their use of the commands. These commands represent fantasies of that heteronormative desire because they offer a restrictive interpretation of pleasure while at the same time presupposing that the participants engaging in the activity have the same response. While this indicates mutual engagement, it also belies a reality in which social actors may have different reactions to the same activity.

Yet, despite the overtly erotic nature of the commands and the sexual narrative contained within them, they do not serve erotic purposes for the community members who use them. The analysis of their function and use shows that while they are a form of erotic mimicry, they serve other ludic purposes too. Their function as games of chance shows how the erotic aspects of the scripts' narratives are distilled into quantifiable interpretations (e.g., duration of sex or breathlessness). This schema for interpretation encourages the participants' reciprocal use of the commands even more than the erotic might. Here, it is possible to observe the social uses of the erotic in order to understand distinctions between articulations of the erotic and those of desire. The analysis of these commands shows how the overtly erotic so often resides within overarching sociosexual discourses and that this occurs even when there are alternative functions and communication goals of it.

The Presentation of Self in Everyday Life

Having Sex" as a Graded and Hierarchical Construct: A Comparison of Sexual Definitions among Heterosexual and Lesbian Emerging Adults in the UK

Homo Ludens: A Study of the Play-Element in Culture

What Love Is: And What It Could Be

Thinking Sex: Notes for a Radical Theory of the Politics of Sexuality

Let's Talk about Race: Identity, Chatbots, and AI

Automation and Design for Prevention: Fictional Accounts of Misanthropic Agency from the Elevator (Lift) to the Sexbot (Chatbot)