Yesterday was about the heralding of creativity, yet there is a setting where creativity goes overboad and the Washington Post gives us ‘Scarlett Johansson on fake AI-generated sex videos: ‘Nothing can stop someone from cutting and pasting my image’‘ (at https://www.washingtonpost.com/technology/2018/12/31/scarlett-johansson-fake-ai-generated-sex-videos-nothing-can-stop-someone-cutting-pasting-my-image).
It is a lot worse than you think and even some might trivialise it by merely hanging onto the idea of women, mostly celebrities being inserted into porn, the problem is actually huge, not merely because of that application. The issue is that whatever I can creatively dream up, I can make a reality if I have enough images. The problem is that most of this software is free (for now), and the problem is growing on two fields.
In the first field we see not merely exploitation of any woman into porn, nudities or weird situations. The technology is close to perfect enough that it is harder and harder to distinguish the fake from the real, the deep fake is overwhelmingly convincing that the Washington Post comment “what your eyes can see and your ears can hear can no longer be taken for granted in the digital field“, that problem is a lot bigger than you think. Even as this example (at https://www.youtube.com/watch?v=BU9YAHigNx8) is easy enough to spot, the overall quality is very much on the high side. Here money gives authenticity, the better the computer, the smoother the result and the shorter the timespan to make that result. The average laptop will not get you far, yet a high end gaming PC gets you an optional 1080 HD deepfake version that is increasingly hard to distinguish from the real deal.
In the second field we see a failing, a failing to investigate and legalise the optional prosecution of deepfake video. Not merely the porn side of it, in its heart the opportunity to crate identity fraud and give rise to interfering with political, social and private lives will go further and will be the foundation of a lot more hardship. 5G will merely send these high end videos faster and wider on a global domain. How long until we get the videos of random politician (Emmanuel Macron) and random Celebrity (Anna Kendrick), whilst places like the Daily Mirror will give the innuendo “European Politician accused of hot Hollywood encounter” with a deep fake? The issue is not that a place like the Daily Mirror would or would not use that image intentionally maliciously, the issue will soon be that their trained professionals can no longer tell the difference between real and deepfake and there is more than one indicator that that moment could optionally be reached this year.
The third part is that not only is this technological field charging ahead, the legal field that should protect the people can no longer keep up, in addition the freedom of expression that allows for ‘creative alteration’ is actually assisting in what should be regarded as criminal activities. In addition there is a larger failing in the law, McMillan (at https://mcmillan.ca/What-Can-The-Law-Do-About-Deepfake). The working allows for a failing that no one is able to deal with. We see this in: “The tort of appropriation of personality arises where a person attempts to gain an economic advantage by using some aspect of another person’s name, likeness, or personality without that person’s consent. To be successful, the plaintiff must establish that the defendant appropriated his or her persona for economic gain. This is also enshrined in legislation such as the BC Privacy Act, which states that it is a tort to use a likeness, still or moving, including a likeness deliberately disguised to resemble a person “for the purpose of advertising or promoting the sale of, or other trading in, property or services.”” You see, the failing is not seen in what must be established, it is found in what is negated. We think of lone wolves as terrorists, yet there is another one, the political lone wolf. A person not recognised by anyone, but does the bidding to promote the political field (or alternative field like product field) of others.
When that person goes after the next political contender we see no ‘gain an economic advantage‘, we alternatively see no ‘trading in, property or services‘, yet the damage will have been achieved. Even when we look in the US regarding: “Malice in law is the intent, without justification excuse or reason, to commit a wrongful act that will result in harm to another. Malice means the wrongful intention and includes all types of intent that law deems to be wrongful“, to enter in court on the premise of malice and intent versus ‘creativity and freedom of expression‘ becomes a first amendment court case where the political player loses no matter what and if the transgressor is young enough the damages will never ever cover the impact. In this day and age of viral video and social media, one or two of these videos could end any political career within a day and that is merely the top of the iceberg. Larger players can lay waste to all kinds of opposition with a much greater ease by giving rise to confusion and relying on the gullible population to spread the damage.
McMillan gives one part that could give the smallest of coverage in British Columbia. As we are introduced to: “The BCHRT has the authority to provide a compensation order for any lost wages or expenses incurred as a result of the contravention, and to provide a compensation award for “injury to dignity, feelings and self-respect.”“, how could any unemployed individual compensate for damages inflicted? The problem would not be the act, the issue would be to a much larger degree the setting of ‘injury to dignity, feelings and self-respect‘ when it is a public figure, especially when we consider the last 10 years where: ‘less than 20% of the orders made are over $10,000‘, in a setting where so far $75,000 has been the highest award. So tell me, who would not fork over $75,000 in some way to win a high end election? Even as the Canadian system is seemingly a little better than most out there, what they have is nowhere near ready to deal with the larger impact the deepfakes optionally have and that is merely outside of the porn application field. The experts that gathered in the SIGGRAPH 2018 annual conference on computer graphics in Vancouver from August 12th to 16th 2018 already agree that this is going to be an issue in the 2020 US elections, the question is will the law be ready and there is no clear indications that it is.
Even as we see that DARPA has been taking an active interest in finding forensic solutions, the department by Dr. Matt Turek, I also see the flawed impact as we see the optional miss that a high resolution result projected deepfake, then captured by a low res camera will create a version that is either negated completely of will be flagged amongst hundreds of thousands of others and still ends up having a 60/40 chance of passing through. Those numbers are not encouraging and this forensic field is for now in its early infancy giving the deepfake players close of half a decade of free reign on every social media with 5G merely impacting a much larger population, much faster. The overwhelming part is that computing power and high resolution recording equipment is becoming more and more affordable so the pool of non-malicious offender is growing close to exponentially for now. Let’s face it which horny teenager has not dreamt of replacing his prom date with the image of Amber Heard, Bailey Noble or Laura Vandervoort as his desired prom date?
It is for the most the absence of intent and malice that is driving the technology. America has 42 million teenagers, close to 22 million are male, most dreaming of the dream date they can never have, so when there is a software company keying in on close to 50% of those offering a $49.95 solution making their dream come true, that is a market surpassing $1 billion, do you think that this is not happening right now? It is a massive driving force, apart from the sex sells part; it is within us to be a pranker/prankster, to replace ourselves with an edited Batman/Batgirl making ourselves the superhero, others want to be seen with their idols (Frances McDermott/Zack Effron or Bailey Noble/Heidi Klum) depending on our age and spreading that imagined tale on Facebook. At heart it is deceiving (read: pranking) their friends intentionally yet completely absent of malice and for every 50 people that do it for the innocent reason, there will be 1-2 malicious people, yet the overwhelming drive for that software is there and the more that want it, the cheaper the solution and for now that this trial software is often free and it is becoming highly perfect in the result, after which it will soon be sold at an affordable price. The problem is that anything innocent can be perverted and the deepfake technology that made yours truly look like Batman (actually Batman was real, Bruce Wayne looks a lot more like me and a lot less like Ben Affleck) can do the same for any exploiter to look the political target look like an arrested person on the street, two of these viral events can totally impact the next elections. Even as the politician itself is the likely target, making the members of his team (like the spokesperson, the strategist and the writer) targets would be a lot more effective, we will dismiss the political person often out of hand, the people they work with less so, it will create doubt and stop their political engine overnight giving the election away to the other person and with the deepfake field evolving at the speed it does, every political party will need a high end reputation management firm in their corner watching out for these attacks and in addition spend too many resources dealing with these attacks, making the small players no longer a consideration and making the larger players spend a whole chunk of money in different areas impacting their visibility. The lack of law, or more correctly stated the lack of impact that the law currently has, will drag elections and public profiles along in very different directions soon enough.
That part is seen (at https://www.ubermetrics-technologies.com/blog/reputation-management-what-to-do-about-deepfakes/) where we see how Jordan Peele is doing the speech for former President Obama and more important, the fact that this looks authentic enough to fool most republicans into an emotional frenzy, game over and the next democratic president starts in 2020, that is the game now and it is frighteningly indistinguishable from the real deal. Even as we see here the stage of ethics, the issue is not merely consent (it only partially is), we see: “the person whose face is superimposed on deepfakes did not give their consent“, in the batman example, my image is the one superimposed and I am giving permission, it is the other part that is owned by DC Comics, or is that Zack Snyder, or perhaps Ben Affleck. Do you think that their engine is ready for millions of trivial cases, often limited to a ‘cease and desist’ order? Their workforce would not be able to deal with 1% of that workload and in the initial race that viral propulsion was optionally used to their advantage. The issue is more loaded than we think and when the court case comes and I would state in my defence: ‘I was merely expressing myself, fantasising on being the next Batman. I made no commercial gain, merely social visibility for my desired optional career in acting‘, do you think that I will get anything more than an optional slap on the wrist as I was pursuing my dream? With the right lawyer it could optionally be thrown out of court as there would be no visible harm or hardship to Ben Affleck, the case ends soon thereafter.
Even as law firms give us the stage of: “must typically prove that the defendant—the person who uploaded the deepfake, for example—published something that gives a false or misleading impression of the plaintiff in such a way to damage the plaintiff’s reputation or cause them great offense, in such a way that would be highly offensive to a reasonable person“, now consider Hustler Magazine, Inc. v. Falwell, 485 U.S. 46 (1988), there we learn that the first and fourteenth amendment is prohibiting public figures from recovering damages for the tort of intentional infliction of emotional distress (IIED), if the emotional distress was caused by a caricature, parody, or satire of the public figure that a reasonable person would not have interpreted as factual. In that setting the deepfake field is wide open to be used against political figured to a much larger extent, that case was never ready for deepfake and by setting the stage to ‘a caricature, parody, or satire of the public figure‘ we see that the reasonable person becomes malleable to the greater extent making the deepfake field a much more effective political swaying tool than we imagined. I reckon that under this setting Larry Flynt will be laughing on the public stage that his case opened up for till the day he dies.