Men with paraphilic interests and their desire to interact with a sex robot

1.Introduction

Sex robots are no longer just science fiction: recently, their development had accelerated and companies are putting considerable effort in technical, behavioural, and visual innovations (Aoki and Kimura, 2021). Sex robots can be defined as human-like, full-sized, anatomically correct humanoid service robots that use a range of different materials and technologies and whose price range varies widely. What they have in common is being designed for sexual use (Döring and Pöschl, 2018). They can differ in apparent age and appearance (e.g., different eye colour, pubic hair, hair, skin colour, makeup) and have several removable orifices for sexual use. Sex robots differ from their predecessors, passive sex dolls, in applied artificial intelligence, which makes them active and responsive. Synthetic speech and simple motor skills allow them to speak, move, and display a rudimentary personality. Although their range of behaviours is still limited, sex robots are likely to become highly sophisticated in the near future.

Attitudes toward sex robot experience (SRE) in the general population have been explored by several small, nonrepresentative, web-based surveys (Appel et al., 2019; Edirisinghe and Cheok, 2016; Richards et al., 2016; Scheutz and Arnold, 2016; 2017; Szczuka and Krämer, 2016). The mean age of respondents in those studies ranged around 30–35 years. The results varied: for instance, in a US sample of male and female respondents (N = 100; 57 male respondents), 70% of male respondents indicated they would use a sex robot (Scheutz and Arnold, 2016), but only 9% of Malaysian respondents of both sexes (N = 32; 16 male respondents) wanted to entertain this idea (Edirisinghe and Cheok, 2016). In a German study focused only on male respondents (N = 263), 40.3% indicated they would buy a sex robot now or within the next five years (Szczuka and Krämer, 2016). This indicates highly varying rates of acceptance of SRE. In the US sample (Scheutz and Arnold, 2016), men were rather accepting of SRE in most areas of use including paraphilic interests (PI), such as unusual sexual practices (e.g., rough sex or sadistic behaviour), but the use of child-like sex robots was strongly rejected by respondents. In a very large Dutch sample (N = 1,162), de Graaf and Allouch (2016) reported that 22% of respondents thought that robotic sexual partners could decrease the incidence of sexual assault and rape. Furthermore, 17% of participants indicated their belief that robotic sexual partners could sexually assist special target groups, for instance people with social or sexual difficulties.

So far, no empirical study investigated the effects of SRE on their user´s partnerships as well as emotional and sexual functioning. Some evidence is emerging from studies on sex doll use, but it is fairly limited and cannot be generalised (Döring et al., 2020). A study on sex doll owners found that they report more issues in their sexual functioning when the sexual activity involves a human partner, while sex dolls provide them with sexual gratification and stimulation (Valverde, 2012). Other studies likewise reported that sex dolls seem to have a soothing and healing effect on users, offer a viable alternative to having a human partner (Knafo, 2015; Knafo and Bosco, 2017), and can alleviate loneliness (Ciambrone et al., 2017).

There is currently a high clinical demand for evidence-based guidelines regarding interactions with sex robots because it is an issue with implications for the physical, mental, and social wellbeing of their users (Cox-George and Bewley, 2018).

2.The use of sex robots in prevention of sexual offending and clinical management of paraphilic disorders

Within the context of sex robot use, two issues tend to attract the greatest controversy: the use of sex robots to prevent actual sexual crimes – such as violent assault, rape, or sexual abuse of children – and their potential use in clinical management of paraphilias.

Disorders of sexual preference (paraphilic disorders) are defined in the International Classification of Diseases, 11^th revision (World Health Organization, 2018) as ‘persistent, intense patterns of atypical sexual arousal, manifested by sexual thoughts, fantasies, urges, or behaviours, the focus of which involves others whose age or status renders them unwilling or unable to consent and on which the person has acted or by which he or she is markedly distressed’. The presence of paraphilic interest (PI) itself has no moral or legal implications, but acting upon it in a way that includes another person presents a potential risk to public safety.

Individuals with PI in minors who deliberately suppress their sexual urges report that their primary motivation for abstaining from offending behaviour is the awareness of harm it would inflict upon a child as well as potential legal consequences for themselves (Walker, 2017). Identified strategies to manage paraphilic urges include the viewing of images and videos of children depicted in mainstream media or computer-generated images of children popularised by the Japanese culture (Walker, 2017), maintenance of sexual contacts with adults, and adult pornography or fantasising during masturbation (Houtepen et al., 2016). Child-like sex robot (or child-like sex dolls, e.g., a model by Trottla) could be viewed as a legal surrogate of their desired partner which has not only a child-typical body scheme but, unlike the child-like sex dolls, also the basic age-typical behavioural displays.

Sex robots, such as the Roxxxy model by True Companion with ‘Frigid Farrah’ personality, may offer an outlet to individuals with PI in non-consent/violence, that is, people who are sexually aroused by infliction of physical or mental pain or by overcoming the active resistance of others (Harris et al., 2012; Seto et al., 2012). Robots such as the abovementioned model can mimic the basic signals of non-consent but the range of emotions or reactions they can display (e.g., defiance, resistance, and suffering) is currently fairly limited.

A recent study that mapped the rates of acceptance of sex robots by therapists and physicians in Germany, Austria, and Switzerland (N = 72), showed that 30% of approached experts could imagine using sex robots for patients with paraphilic disorders (based on ICD-10), although contrasts between respondents’ opinions on this subject were high. Respondents stressed that the benefits of SRE must be decided individually for each specific patient. For some, it could offer an opportunity to express their sexuality with robot as a substitute, whereas for others, it could intensify their sexual urges and lower barriers to committing a sexual offence (Eichenberg et al., 2019). The survey also reported significant differences in attitudes towards the use of sex robot in therapeutic setting between younger and older therapists, with younger therapists being more open to the subject (Eichenberg et al., 2019).

Parallel information regarding the acceptance of and desire for SRE among potential users is largely missing. The goal of this study is to fill in this knowledge gap: after all, willingness to engage with sex robots would be a necessary prerequisite of any potential therapeutic use. Our aim is to explore the level of desire to engage in SRE in individuals with PI, particularly those for whom finding a desired partner is unrealistic and/or illegal (e.g., individuals with PI in minors) or at least difficult and mostly unsatisfying because their partners’ sexual behaviour is not complementary to their PI (e.g., individuals with PI in non-consent and/or sexual violence).

First, we explore the percentage of men in a representative sample of Czech population interested in SRE and the percentage of men with PI in minors and in non-consent/violence interested in SRE. We focus solely on men because paraphilic interests seem far more prevalent amongst men than women (Bártová et al., 2021; Bouchard et al., 2017). Secondly, we investigate the differences in SRE desire between the abovementioned samples. In particular, we hypothesise that due to the impossibility to fully express their innate sexual desires with actual people without harming a child/adult and breaking the law, both of these groups with PI would express greater desire in SRE than men in the general population (control group). Furthermore, we expect that younger participants would be more open to the idea of SRE due to their generally greater acceptance of new technologies (Eichenberg et al., 2019), which is why we will control for the effect of age in a complex model which analyses the differences in the level of desire for SRE across the groups.

3.Methods

The prevalence of SRE desire in the general population was assessed using a sample of Czech men engaged in an online representative survey on ‘Love and intimacy in the Czech population’. The sampling was conducted by a respected sociodemographic agency. Separately, a sample of men who met the criteria of PI in minors and PI in non-consent/violence was pre-screened from an online pool of Czech respondents to assess the prevalence of SRE in the two target groups. All participants completed a questionnaire which contained a question regarding their desire for SRE. They expressed their level of agreement with the following statement ‘I desire sexual experience with a sex robot/doll.’ using a 4-category ordinal scale (strongly disagree, disagree, agree, strongly agree). The study was approved by the Ethics Committee of the National Institute of Mental Health under the Nr. 41/15.

4.Results

A detailed summary of descriptive parameters of the groups is presented in Table 2.

SRE desire was expressed by 18%, 37.5%, and 26% of respondents in the group control, group with PI in minors, and group with PS in non-consent/violence, respectively (Table 3).

Across the three groups, we found a significant difference in the distribution of binary SRE desire (X2(2)=13.256; p=0.001; Cramer’s V = .121). The control group and group with PI in minors differed significantly in expressed binary SRE desire (X2(1)=11.56; p=0.001; Phi φ=−.116). Between the control group and group with PI in non-consent/violence, expressed desire for SRE did not significantly differ (X2(1)=2,617; p=0.106; Phi φ=−.055).

Categorical regression (CATREG) model explained nearly 7% of variation in SRE desire scale (R2=.069; Radj2=.065). The model was statistically significant (F(3,907)=22.25; p<.001) and SRE desire scale was significantly predicted by both group membership (β=.159; F(2)=20.17; p<.001) and age (β=−.195; F(1)=38.48; p<.001).

Multiple Mann–Whitney (M–W) U tests for pairwise comparisons revealed that the group with PI in minors (N = 48) differed from the control group (N = 806): M–W U = 12,577.00; p<.001; r = .16 (median_minor = ‘disagree’, median_control = ‘completely disagree’). The situation was parallel for a comparison between the group with PI in non-consent/violence (N = 57) and the control group: M–W U = 18,286.50; p<.01; r = .10 (median_{non-consent/violence} = ‘disagree’; median_control = ‘completely disagree’; see Fig. 1). Nevertheless, neither of the differences was practically significant because the size of the effects did not reach the recommended minimum effect size for a ‘practically’ significant effect in the context of social science (Ferguson, 2009).

Visual explorations revealed that younger respondents tended to be more accepting of the idea of sex with a robot, but the size of this effect only reached the recommended minimum effect size for a ‘practically’ significant effect (Ferguson, 2009): Spearman’s ρ=.204, p<.001, N = 911, Rs2=.042.

Fig. 1.

A violin plot showing expressed SRE desire rated on a four-point scale (1 – completely agree, 2 – agree, 3 – disagree, 4 – completely disagree) by groups.

5.Discussion

Our study explored desire for a sex robot experience (SRE) in three groups of male respondents: first, a general sample of the Czech population, second, individuals with PI in minors, and third, individuals with PI in non-consent/violence. We expected both target groups (with PI in minors and with PI in non-consent/violence) to have a greater desire for SRE than the control group. We also expected that younger individuals would express greater SRE desire. The proportion of interested individuals ranged from 18% to 37.5% across groups. Statistical differences in the desire for SRE were found between the control group and the group with PI in minors but not between the control group and group with PI in non-consent/violence. Differences between the control group vs. group with PI in minors and the control group vs. group with PI in non-consent/violence in the desire for SRE were calculated using a model controlled for the effect of age. Differences between the groups were not practically significant (Ferguson, 2009) and the effect of age reached only the recommended minimum effect size (Ferguson, 2009).

The control sample expressed the least desire in SRE of all groups (18%). The desire for SRE in our control group was lower than that of respondents from a general US population sample (70%; Scheutz and Arnold, 2016) but higher than in the Malaysian sample (9%; Edirisinghe and Cheok, 2016). It should be noted, though, that studies by Scheutz and Arnold (2016) and Edirisinghe and Cheok (2016) used small convenience samples (N = 100, N = 32 respectively), whereas our control group was a representative sample of Czech male population (N = 806) recruited online from a national pool of respondents. Other reasons for different levels of desire for SRE may include differences in the cultural background, which could affect openness to new technological advances, or the mean age of respondents.

Although differences in the desire for SRE did not reach the level of practical significance (Ferguson, 2009) across the groups, we found that a relatively greater desire for SRE was expressed by the group with PI in minors. This could be attributed to the fact that persons with PI in minors cannot fully express their innate sexual desires without violating the law, which might motivate them to look for alternative routes of sexual expression. A sex robot might be perceived as a strategy of managing their PI.

Individuals with PI in non-consent/violence also expressed a greater desire for SRE than the control group. This might may be associated with either their general greater interest in impersonal sexual encounters, reflected in their greater use of commercial sex services (Farley et al., 2011; 2017), or with perceiving SRE as a way to satisfying one’s PI without inflicting harm by sexual assault or rape.

The minimal effect of age on SRE desire was surprising because younger individuals are thought to be more accepting of new technologies and novel sexual experiences. Along similar lines, Eichenberg et al. (2019) reported that younger therapists were more open regarding a possible use of sex robots as part of therapy than older therapist were. On the other hand, Scheutz and Arnold (2016) reported somewhat different results: in their study, younger people did not express a greater acceptance of sex robots than older respondents. On the contrary, younger people were less accepting of the idea of interacting with sex robots rather than with prostitutes than older people were Scheutz and Arnold (2016). It might be due to the fact that younger generations are sensitised to issues pertaining to consent and ethical sexual conduct.

6.Limitations and suggestions for a future research

The study explored the prevalence of desire for SRE across three groups. The results should be interpreted cautiously for the following reasons: First of all, recruitment strategies for the samples and the numbers of respondents across groups differed, with the control group being distinctly larger than either of the PI groups. This is a consequence of the way the target groups were defined (since our aim was to recruit people with a potential medical diagnosis): by their nature, samples of people with PI cannot be representative. Secondly, the study did not explore respondents’ preferences regarding sex robot’s features (apparent age, physical appearance, or behavioural displays). We can only speculate whether such preferences would indeed reflect, in the relevant groups, the nature of the PI. Instead, we only assume that people with PI in minors would prefer child-like robots with age-typical behavioural displays and people with PI in non-consent/violence would choose adult-like sex robots with clear displays of resistance or strong negative emotions. Thirdly, the survey did not include a question whether respondents would welcome a sex robot as part of their therapy, and it should be noted that expressed SRE desire is not the same as interest in using a sex robot as part of therapy. And finally, the survey insufficiently differentiated between the terms ‘sex robot’ and ‘sex doll’ when asking about respondents’ interests: the statement in point was worded as ‘I desire sexual experience with a sex robot/doll.’ Sex dolls may be more familiar to respondents because they have been on the market much longer than sex robots and they might have a connotation of passivity. As a consequence, respondents may have failed to consider the sex robots’ interactive features.

These shortcomings should be considered and addressed in future research to help us better understand the potential of sex robots in therapeutic settings. Finally, as noted above, the present study did not aspire to solving the question of whether sex robots are a suitable therapeutic tool for people with PI or whether they could play a role in the potentiation/suppression of sexual offending behaviour.

Exploration of desire for SRE within the clinical population is a prerequisite for the currently extensively developing discussions about the use of sex robots in clinical management of paraphilic disorders. Our study is the first to provide data in this direction.

Funding

This study was supported by the Institutional Support for Longterm Development of Research Organizations – Charles University, Faculty of Humanities (Charles Univ, Fac Human 2021), by Specific University Research 2021 – 260 610, by Charles University Research Centre of the Charles University, Czech Republic (UNCE 204056), and by project ‘Sustainability for the National Institute of Mental Health’ under grant no. LO1611, with financial support from the Ministry of Education, Youth, and Sports of the Czech Republic under the NPU I program.