| |
| META TOPICPARENT | name="FirstEssay" |
|---|
It is strongly recommended that you include your outline in the body of your essay by using the outline as section titles. The headings below are there to remind you how section and subsection titles are formatted. |
|
<
< | Deepfakes and Our Right to Sexual Privacy |
>
> | Deepfakes and Sexual Privacy Laws |
| | -- By KerimAksoy - 05 Oct 2019
Introduction |
|
<
< | In just over two months, around 6 million people have watched on YouTube? the deepfake video of Bill Hader channeling Tom Cruise. Doing so is an unnerving experience. Hader, who is known for his particularly uncanny impressions, seamlessly changes faces as he pretends to be Tom Cruise. The transformation is so seamless, in fact, that it is likely that most people would be fooled into believing the person sitting in front of David Letterman is Cruise and not Hader. In this case, deepfake technology was used for a humorous, not sinister, purpose. It was clearly labeled as a deepfake video, and it depicts an interview already available online. However, deepfake technology poses serious threats to our right to privacy, especially our right to sexual privacy, and the law is not sufficiently equipped to combat it. |
>
> | Millions have watched the deepfake video of Bill Hader channeling Tom Cruise. Hader, who is known for his uncanny celebrity impressions, seamlessly changes faces as he pretends to be Cruise. It is likely that most people would be fooled into believing the person sitting in front of David Letterman is Cruise, not Hader. In this case, deepfake technology was used for a humorous purpose. It was clearly labeled as a deepfake video, and it depicts an interview already available online. However, deepfake technology poses serious threats to our right to sexual privacy, and the law is not sufficiently equipped to combat it. |
| | What Are Deepfakes?
Definition |
|
<
< | The term deepfake is a portmanteau of "deep learning" and "fake." Such videos can be created using a Generative Adversarial Network (GAN), which is a machine learning technique that was invented to "algorithmically generate new types of data out of existing data sets." For example, this software can be fed thousands of bits and pieces of melodies or thousands of pictures of human beings, and it can then create its own song or picture of a human being that is distinct from the data sets it was fed. In the case of videos, for example, this software can be used to superimpose the face of one person onto the body of another. |
>
> | The term deepfake is a portmanteau of "deep learning" and "fake." Such videos can be created using a Generative Adversarial Network (GAN), which is a machine learning technique invented to "algorithmically generate new types of data out of existing data sets." This software can be fed thousands of pictures of human beings, and it then creates its own picture that is distinct from the data sets it was fed. |
| | Brief History |
|
<
< | Ian Goodfellow, a graduate student, created the GAN technique in 2014 for research purposes, and it was not until late 2017 that the technology started becoming more mainstream. It was then that a Redditor by the name of "Deepfakes" started posting videos of female celebrities superimposed onto the bodies of pornographic actresses. "Deepfakes" used TensorFlow, Google's open-source machine-learning software, to build GANs with such capabilities. Other Redditors joined suit, and their Reddit thread attracted enough attention to warrant media coverage. In response, Reddit banned these users, but the cat was already out of the bag. "Deepfakes" created a free software called "FakeApp," which lets anyone easily create their own fake videos. At this point, deepfake technology was democratized - for better or for worse. |
>
> | Ian Goodfellow created the GAN technique in 2014 for research purposes, and it was not until late 2017 that the technology went mainstream. It was then that Redditor "Deepfakes" started posting videos of female celebrities superimposed onto the bodies of pornographic actresses. "Deepfakes" used TensorFlow, Google's open-source machine-learning software, to build GANs. Other Redditors joined suit, and their thread attracted significant media coverage. Reddit banned them, but the cat was already out of the bag. "Deepfakes" created a free software called "FakeApp," which lets anyone easily create their own fake videos. Deepfake technology was democratized - for better or for worse. |
| | |
|
<
< | The Law of Sexual Privacy |
>
> | Laws of Sexual Privacy |
| | Tort Law |
|
<
< | There are a number of provisions in tort law, as delineated in the Restatement (Second) of Torts, that are applicable in analyzing the sexual privacy implications of deepfakes. The best rule is 652(e)- Publicity Placing Person in False Light. This provision states that a defendant is liable to a plaintiff for placing the plaintiff in a false light if: (a) "the false light in which the other was placed would be highly offensive to a reasonable person," and (b) "the actor had knowledge of or acted in reckless disregard as to the falsity of the publicized matter and the false light in which the other would be placed." Another highly applicable rule is 652(c) - Appropriation of Name or Likeness. This provision provides that "One who appropriates to his own use or benefit the name or likeness of another is subject to liability to the other for invasion of his privacy." |
>
> | There are two rules in tort law that are applicable to combating the sexual privacy implications of deepfakes. The best rule is 652(e)- Publicity Placing Person in False Light. A defendant is liable to a plaintiff for placing her in a false light if: (a) "the false light in which the other was placed would be highly offensive to a reasonable person," and (b) "the actor had knowledge of or acted in reckless disregard as to the falsity of the publicized matter and the false light in which the other would be placed." Another highly applicable rule is 652(c) - Appropriation of Name or Likeness. This rule states that "One who appropriates to his own use or benefit the name or likeness of another is subject to liability to the other for invasion of his privacy." |
| | State and Federal Laws |
|
<
< | In response to a rise in revenge porn websites, 46 states and the District of Columbia have passed laws criminalizing the publication of nonconsensual pornography. So far, Massachusetts, Mississippi, South Carolina, and Wyoming are the remaining holdouts. New York's law, for example, makes it a misdemeanor to unlawfully disseminate or publish an intimate image. The perpetrator must have had the intent to cause harm, s/he must have had the intent to disseminate or publish, and the victim must be identifiable and have an intimate body part exposed, among other elements. California recently passed a law that goes further and allows a "person edited into sexually explicit material without consent to seek civil damages against the person who created or disclosed it." There are also new federal laws in the works that would "penalize distribution of malicious forgeries and impersonations, including deepfakes." |
>
> | 46 states and the District of Columbia have passed laws criminalizing the publication of nonconsensual pornography. New York's law, for example, makes it a misdemeanor to unlawfully disseminate or publish an intimate image. The perpetrator must have had the intent to cause harm and the victim must be identifiable and have an intimate body part exposed, among other elements. California recently passed a law that goes further and allows a "person edited into sexually explicit material without consent to seek civil damages against the person who created or disclosed it." There are also federal laws in the works that would penalize the distribution and creation of malicious forgeries and impersonations. |
| | Possible Solutions
Tort Law as a Solution |
|
<
< | The false light tort can be a strong tool in helping victims sue for damages when their faces are superimposed onto the bodies of performers in sexually explicit videos. It is admittedly reasonable to believe that a jury would find that both elements of the false light tort would be met in such cases. The first element, that the false light in which the plaintiff is put would be highly offensive to a reasonable person, is tied to changing cultural norms. Regardless, it goes without saying that most individuals, when surveyed, would find it highly offensive that a fake video of them engaging in sexual acts - or any video, even if real - is published. The second element would likely easily be met as well, since anyone creating a deepfake would do so knowingly and by mistake. However, it is another matter as to whether they knew it to be fake when simply sharing it and not creating it. Regardless, however, even if they believed it to be real, sharing such a video would put the person in a false light and they would still be liable. |
>
> | The false light tort can be a strong tool in helping victims when their faces are superimposed onto the bodies of performers in sexually explicit videos. In regards to the first element, most jury members would find it highly offensive when a fake video of them engaging in sexual acts is published. The second element would likely be met as well, since anyone creating a deepfake would do so knowingly. However, it is another matter as to whether they knew it to be fake when simply sharing it and not creating it. Regardless, however, even if they believed it to be real, sharing such a video would put the person in a false light and they would still be liable. |
| | |
|
<
< | The misappropriation tort is another strong tool that gives individuals control over their image, although in the deepfake context, it's not without its limits. The misappropriation needs to be for the defendant's use or benefit, and some states limit such use or benefit to commercial use. In such states, victims would be unable to sue under this tort if the perpetrator simply shares the deepfake video online and does not make any money from it. Incidental use can also be a defense, When you show a bunch of people, it’s hard for any one person to claim their specific identity was benefitted from as opposed to the group’s. |
>
> | The misappropriation tort is another strong tool that gives individuals control over their image, although in the deepfake context, it is not without its limits. The misappropriation needs to be for the defendant's use or benefit, and some states limit this to commercial use. Victims would be unable to sue under this tort if the perpetrator simply shares the deepfake video online and does not make any money from it. Incidental use can also be a defense to a claim of misappropriation. When you show a number of people, it is hard for any one person to claim their specific identity was benefitted from as opposed to the group's. |
| | Revenge Porn and Related Laws as a Solution |
|
<
< | Revenge porn laws are unfortunately a weak solution to the problem of deepfakes. Take the New York law, for example. |
>
> | Revenge porn laws are not the strongest solution. One of the elements of the crime in the New York law is that the "dissemination must depict an unclothed or exposed intimate part of such other person..." Whether deepfake videos depict an "unclothed or exposed" part of the victim is a matter of interpretation. A court could rule that, since a deepfake video shows the body of a pornographic actress. It is not the victim's intimate parts that are being disseminated, but that of the pornographic actress.
On the other hand, laws that go further by criminalizing the distribution of "malicious forgeries" are much more likely to be effective and serve as a deterrent to the creation and distribution of deepfakes. They are aimed squarely at what deepfakes are - forgeries - and they do not specify whether the victim's intimate parts need to be shown. Their analysis simply rests on the video being fake. |
| | Conclusion |
|
>
> | Deepfakes subvert our sense of reality and could lead us to further lose faith in video as a "reliable record of reality." The law offers only partial remedies to the problem. In order to combat it fully, the federal government and all states need to pass laws targeting the creation and dissemination of malicious forgeries. |
| |
You are entitled to restrict access to your paper if you want to. But we all derive immense benefit from reading one another's work, and I hope you won't feel the need unless the subject matter is personal and its disclosure would be harmful or undesirable.
To restrict access to your paper simply delete the "#" character on the next two lines: |