Forbidden Research liveblog: Sexual deviance: can technology protect our children? | MIT Center for Civic Media

Forbidden Research liveblog: Sexual deviance: can technology protect our children?

liveblog by Alexis, Sam Klein, Natalie, and myself

Ethan Zuckerman, Director, MIT Center for Civic Media moderates.

Conducting research on adults who have sex with children is virtually impossible due to ethical and legal restrictions. The advancement of technologies like robots and virtual reality has opened the door to exploring questions that were previously not possible. But while a U.S. court case has held that virtual child pornography is legal, the law in this area is controversial and emotionally charged. Legal uncertainties and vast stigma make actual research difficult. At the same time, a better understanding of this deviant behavior has the potential to significantly change lives.

Lead to paraphilia. We're not showing explicit imagery. It may be triggering. Going to try to deal with this very difficult topic. Lots of real world ramification. 1300 people are serving time for sex crimes just in MA. X are in indefinite civil confinement - finished sentence, but not released into the general public because of fear of recidivism. Research on the statistics. 10% to 50% which suggests that there isn't a ton of research. We know very little indeed. Most people who are afflicted with pedophilia are actively trying to fight these urges. When talking with therapists, they're trying very hard not to act on these urges they're suffering from. There are some existing efforts to develop support programs. Whether there are ways of treating with VR, intimate robotics, etc.. It's a challenging topic with a lovely set of folk willing to take it on.

Kate Darling, MIT Media Lab, IP Theory, Policy, and Robot Ethics, Fellow at Harvard Berkman Center is looking at the NOW of robot-human interaction. Leads us off. Human robot interaction, how we behave around robots. People treat them as though they're alive. We know they're just machines, but subconsciously when we interact, we treat them as if they're alive. Not just about people getting used to a new technology but instead something which is biological. Our brains might project intent and life on moving things which seem autonomous. How strongly we respond to the cues these machines give us. Gives us a chance to study human psychology. People who have low empathic concern for others treat robots differently than those who have high empathetic concern —this is part of a research study Kate has been conducting with Palash Nandy, a researcher at the Media Lab in the Personal Robotics group. Those who treat robots like a living thing makes them a potentially great tool for

When child-size robots come to market, will they be used to address desires and protect children, or to normalize it and put more children at risk? There's no way for us to know. These urges are not a moral failing, they are a psychological issue. Nearly impossible to self-report as you'll get booked. If we really care about children, we might need to be preemptive about this.

Courts don't know what to do with these robots, since no child has been harmed in making them.
While high quality sex robots are not coming as quickly as some might think (or like), but they are coming at a pace that's faster than society is willing to talk about.

Child porn doesn't exist for at least two reasons: because we think it's not ok, but also because a child was harmed in its creation. 3D modeling etc would shift that. Do any international courts handle this differently? Need intent as well as harm to have broken a criminal law in the US. In the US in '96 we had an act forbidding pornographic CGI depicting children. In '02 the supreme court decided there was a free speech issue that overruled, and struck down parts of that act. Since then, a new act "Protect" has been passed, which prohibits "obscene" cartoons. They've shifted the child piece towards obscenity, which is not well defined and depends on community standards. Ex: some media showing young girls performing fellatio was targeted b/c it was aimed at an audience of young girls to teach them improper behavior, not because it included images of young girls.

Back to the question of legal status, but what is exploitative and what is put into legal frameworks because it is uncomfortable?

Ron Arkin, Roboethicist and Professor, School of Interactive Computing, Georgia Tech. Robots and robot ethics. Sex, laws, and violence. Robot deception, killer robots, but today we're talking about sex.

There's plenty of money to be made in lethal autonomous weapon systems. The USG doesn't do? that, but some people do. And I do work in robot deception (how and when robots lie?) but that's not what we're discussing today. Outbranch from Genevive Bell. Intimate Robotics. Most concerned with lethal autonomous systems, but also concerned about something else that is happening now. I work with Sony, and the Aibo. we know how to make people fall in love with these things. What if we start crossing from teh social space into the sexual space? the questions we ask are around if human-robot intimacy is acceptable? Do you become a deviant if you have sex with a robot? [What about one indistinguishable from a human?]

Can [sex with a robot] it serve sort of like methadone for [sexual] deviants? These are research questions which need to be explored. Any time past offenders are released back into society, there will be more victims. [as there are already today] We need to be prepared for that.

Thank you for this forum for sharing this discussion with the audience and the public.
Uncanny valley has to do with behavior, temperature, texture, etc. Roxxxy, VR robot. How will the prostetution industry do in light of these dolls etc? Some see it as a way to free prostitutes to do other things (this is a rather paternalistic view). [Examples from non-sex robots, and sex robots, the former already well-funded by governments.]

Article from the Atlantic recently: "Can Child Dolls Keep Pedophiles from Offending?"
VICE: "Canada's Child Sex Doll Trial Raises Uncomfortable Questions About Pedophilia and the Law" A man ordered one in Canada and gets arrested. Methadone is perscription, maybe we need to do the same here.

Protest: there's a group called Campaign Against Sex Robots, viewed by the founder as "part of a cultural pattern to legitimate pedophilia more widely". [This reminds me so much of some of the arguments against prostitutions, not seeing how sex is a basic human need ].

Study from Stanford University — touching a robot's 'intimate parts' makes people uncomfortable. Published in an obscure journal, couldn't find funding source listed. Study on body comfort zones: differences between Japanese and American adults re: where they normally touched close friends or sexual partners, and where they touched their parents, when they met. Dramatic differences in all cases; which one would take into consideration with any robot designs.

Wanting to establish a research agenda. Will sex robots increase or decrease urges? We have an ethical obligation to do this research.

Christina Couch, Journalist
Written on the question of computer imagery. How our feelings, thoughts, desires shape our our technology is designed.

Working on an article - therapeutic uses of VR : Treating PTSD, depression, phobias, addictions. Patrice Reneaud studying pedophiles right at the point of arousal. Having someone at that point requires a stimuli, which for pedophiles have super valid concerns attached to them.

Dr. Patrice Renaud - "assessing deviant preferences in sexual offenders" using virtual immersion.

Mostly audio, having a hard time getting data. His team published a paper that even when they KNOW a subject is a known pedophile, they couldn't evoke a response just using audio files. When VR started working, they started building scenarios with [fewer] concerns. The differences are in motor and eye movement. The data we have on recidivism is nebulous because it's really hard to study this group. Interested that VR is what opened up these possible studies.

Also a shift in how we view pedophiles. Virtuous Pedophiles is a support group of people trying to prevent acting on their impulses. Another group like them is called the Dukelfeld Project. Confidential treatment focused on preventing them from acting out. Accessibility to a population which has been traditionally hard to research PLUS new research tools usually means a huge amount of research coming out, but not so much for this case. Dukelfeld was able to get on their feet because there aren't reporting issues, it might be covered by insurance.

In the US, Lupron (what people use to chemically castrate themselves) can be gotten under perscription as well. We'd first need a better idea of this as a psychological problem rather than a moral problem. How is the legal system changing? We used to think homosexuality was a moral failing, and then a disease, and is now accepted. How does that transition happen? Popular culture led. TV shows etc. The law came afterwards. Untouchable was a documentary on this. If you throw the word "pedophilia" into any legal debate, all the politicians jump in to vote for it. There's no nuance. The topic (understandably) raises so many emotions for folk that it's difficult to have a rational conversation about it. The YouTube videos for NYT articles are saying they need to be slaughtered for saying it might be an illness rather than a moral feeling. European culture just thinks about porn files as any other file. It allows them to approach all of this in different ways.


So: you're a robot ethicist. What's stopping you from studying this? A: Funding. It can come most easily from foundations, but only a limited amount does. Most governments don't support it [though as noted elsewhere: in Germany treatment for pedophilia is covered by insurance].

Can people get help with their urges? What are their motivations for doing this work? Is there a trend in what's going on?
I don't know that the landscape of research today is big enough to generalize. I know of only 2 researcher using VR for this. Maybe there's more than that, but it is small.

Q: What are the key research questions in this space (intimate robotics)? As an experimentalist, are there things you're thinking about studying that will help answer these Q's?
A: (Kate):- how do our interactions with robots affect our interactions with humans? I think about harm: if our interactions with robots lead to harmful interactions with other humans, than that is a bad thing. But we don't know if our interactions with robots will lead to any of these negative outcomes, so we have to do the research.
We currently know how to observe and measure behavior with robots, but we don't know if this changes your behavior with people. That's hard to study in any arena, and especially with intimate robots, I'm worried we won't do it at all. If a big name in social robotics can't get support for such work, how can anyone do that?

Q: In defense of continental Europe and its culture - I have experience with Quaker work in this field; they are some of the frist to visit prisoners, and they work with sex offenders. All these technologies seem to further sequester and isolate offenders from society. That seems to be something they're trying hard to get over; they want to integrate with society. Do you think it is possible in the US to propose a system where groups voluntarily engage w/ pedophiles, on the basis that they are humans, and shouldn't be treated even worse than violent criminals?
A: (Christina) A story just came out about a [NJ?] task force that is dedicated to that. I don't have any way to answer, but people are starting to consider it.
A: (Ron) I don't think this is a panacea for reintegration; but it should be considered as a possible positive force.
A: (Kate) People being visited are the small percentage who have been convicted of a crime. Most people affected have a hard time coming forward at all. I love that groups are forming to help (convicts), but we need legal changes and maybe tech changes as well if those can help people.

Q: (Willow) You're talking about a lack of data and other things. This is the most academic panel we've had so far. I wonder if by opening up science to more citizen science approaches like other panels have, we might catch more data and discover more things. Are there possibilities to try citizen science approaches to sex studies?
A: (Ron) Yes, and I support anyone who wants to contribute to this, but: this does need strictly controlled science evaluation, with IRB and other controls, to get reliable data. We couldn't even get accurate recidivism rates, and how hard could that be? Numbers were all over the place, because of the dearth of data. And we need to understand the tech coming down the pike. That might be easier to distribute, rather than the study of sexual deviance.
A: (Christina) VR is also bing used to study and treat victims of sexual trauma. This (type of tech) isn't a one-way street. When we talk about amplifying research methods, it's not just for offenders.

Q: You all seem to say that lack of funding and stigma are barriers. What would this look like if those didn't exist? Get to Ethan's initial questions: how does culture change, this medium change, over the long term? How does this change the future of dealing with taboo paraphilias?
A: (Ron) Human research interaction work has at present gone almost too far in requiring huge amounts of data. I think we can develop research with small focused sets of data. Then a series of progressive experiments could gather a larger body of data over time.

Q: There are studies that suggest porn has changed standards of sex. If these child robots are rolled out, who gets to decide if you can have such robots as therapy, or as entertainment?
A: (Kate) A lot of research here has been very [basic]. Many studies have questioned if violent/sexual games change behavior. Methods used in those cases can be applied to robotics. The increased realism may have more of an effect, but similar approaches may apply.
A: (Ron) To my mind, courts or physicians would say it is appropriate. But there would have to be quasi-controlled environments for this to work.

Q: (Victoria) I'm interested in areas where sex is considered? inappropriate such as w/disability. Looking at sexual deviance - how much correlation or research has been done to see whether pedophilia is different from other deviances? (anything considered unnatural by any society: bdsm, homosexuality, &c) If we're talking about changing human behavior, can we cross-examine sth like violence, and the relevant effects of VR? [To inform these questions in a less taboo realm.]
A: (Christina) I don't know if any correlation like this has been done. In 20y of VR research, we find that yes it can influence RL behavior. PTSD can be reduced more quickly... unconscious racial bias can be reduced (in at least one study), though that's incredibly difficult to break even temporarily. It's not crazy to think this could influence behavior in pedophiles.
A: (Ron) Changing human behavior happens routinely, with or without robots - a change wrt. views on marriage, bestiality

Willow question: why can't people consent to sexual research?
Question: how do advanced robots consent to sexual research?
Willow question: why are we talking about fully autonomous systems? Does the debate in self-driving cars and mixed-control systems apply here?