TikTok Tom and the Ethicality of Deepfakes

I was scrolling through my TikTok feed one day when I came across a video that made my finger pause from its continuous swiping motion. The video featured Hollywood superstar Tom Cruise–or someone (or should I say something) who appeared to be Tom Cruise–teeing off at a golf course. My curiosity getting the better of me, I decided to tap on the user’s @ to investigate whether or not the mega-celebrity had really created a personal TikTok account. I found myself quickly becoming a part of the new viral intrigue that was TikTok Tom. The comment sections were flooded with zingers: “Day 173 of not knowing whether this is the real Tom Cruise or not…”, “He looks more like Tom Cruise than Tom Cruise does,” and my personal favorite, “If you are being held against your will, blink twice.”

Unfortunately for one hopeful fan who insisted that “That’s Tom Cruise because he has crooked front teeth which are off-center a little bit to the left,” it wasn’t really Cruise. Instead, the videos featured a deepfake doppelganger of Cruise created by visual and AI effects artist Chris Umé with the help of actor Miles Fisher. From the slightly misaligned front teeth to the classic action-hero gait, the deepfake was incredibly convincing–and, as one viewer put it, “worryingly good.” While more and more social media users gravely predicted that the TikToks were a chilling sign of an apocalyptic future, TikTok Tom was setting off similar panic in national security and intelligence circles. Ever since the first deepfakes debuted in 2016, the technology has been viewed as an emerging threat to national security, personal identity, and even truth, itself. As the Tom Cruise videos have shown, that threat is more prominent than ever, as technology is now so advanced that virtually any person with the right equipment can produce lifelike fakes.

Thus arises an important question: is the use of deepfakes ethical?

The vast majority of people seem to find the idea of deepfakes discomforting. There’s just something inherently morally problematic about footage that’s able to realistically portray people saying or doing things that they’ve actually neither said nor done. It would, however, be unproductive to be so quick to condemn an entire technology without considering all of its facets. The question I have presented is extremely broad, after all, and it hugely depends on individual circumstances. For instance, fans of the Star Wars franchise were delighted, rather than disturbed, when YouTuber Shamook created a deepfake rendition of Luke Skywalker that heavily improved his cameo in The Mandalorian.

I believe that it is difficult to confidently say that a use of a deepfake is unethical without examining the ACM code of ethics and considering 3 additional central factors, as proposed by Adrienne de Ruiter in “The Distinct Wrong of Deepfakes.” They are as follows:

1. Would the deepfaked subject object to the way that they are being represented?
2. Does the deepfake deceive viewers?
3. What was the original intent of the deepfake?

In this essay, I’ll be analyzing a specific case: TikTok Tom.

After his TikToks inspired mass panic on the Internet, Ume had to go on a media blitz to quell some of the hysteria, revealing how he made the videos and reassuring the world his intentions were only to entertain–not to spread misinformation.

However, following the approach I have outlined above, I argue that Ume’s Tom Cruise deepfake, while made with no malicious intent, is unethical.

As previously mentioned, Ume explicitly states that his only purpose for creating the deepfakes was to entertain the public. The videos contain footage of “Tom Cruise” playing the guitar and marveling at a bubblegum lollipop, content that steers clear of damaging and grossly misrepresenting Tom’s character. So far, the TikToks appear to honor all of the relevant ethical principles laid out by the ACM Code of Ethics. While it’s difficult to say that the videos “benefit society” (1.1), they do not cause harm (1.2); actively seek deception (1.3), especially given that Ume has publicly been more than transparent about the deepfake nature of his videos, even posting video breakdowns that showcase some aspects of how he creates his deepfakes (though there will always be confused viewers who forget to read the username @deeptomcruise); nor disrespect privacy (1.6), as Ume’s computer model only uses images of the actor available to the public domain. What makes Ume’s usage of deepfakes unethical, then, is what I believe to be a violation of central factor 1: the aspect of consent. Ume stated in an interview that his team has reached out to Cruise’s management, asking if they should take the videos down, but they never received a reply. Given that some action (legal or not) would have been taken by Cruise’s management team had the videos been perceived as even slightly damaging, some may interpret the lack of objection as consent. However, in this new age of deepfakes, I argue that protection against the manipulation of realistic digital representations of our images and voices should be a fundamental human right that is carefully preserved, and that entails defining consent not as the lack of disapproval but as the clear expression of approval from the person being represented. In accordance, the lack of explicit consent from Tom Cruise means that Ume’s deepfakes are in violation of Tom Cruise’s right to autonomy in influencing the way he is digitally represented, and is, therefore, a violation of Code 1.1 of the Code of Ethics.

The uses of deepfakes (benign, neutral, or unethical) are vast and varied. Sadly, the most common current use of deepfake technology is for pornographic purposes. There are further concerns that deepfakes will be used in the near future to “undermine trust in the democratic process and institutions, heighten social and political tensions, commit crime, destabilise financial markets, subvert diplomatic relations, and incite violence” (de Ruiter, 2021). It is important to note, however, that the technology has been used for beneficial and ethical purposes, as well, such as in the development of software that can artificially regenerate the voices of people who are unable to speak due to illnesses like ALS. Rather than censor the technology, it is imperative that we develop deepfake threat models, ethical AI principles (similar to the ones I have outlined above), and regulations that promote awareness, encourage advancement, and do not stifle innovation. Even Chris Ume has said that he strongly believes that there should be laws that help with the responsible use of deepfakes. Though I have ultimately concluded that Ume’s use of a deepfake was unethical, it is undeniable that there is a silver-lining to it all. In his own words, “I think it’s a good thing I created these videos because now I’m raising awareness and they realise, ‘This is real. It’s coming.'”

References

[1] Corcoran, M., & Henry, M. (2021, June 23). The Tom Cruise deepfake that set off ‘terror’ in the heart of Washington DC. ABC News. Retrieved October 22, 2021, from https://www.abc.net.au/news/2021-06-24/tom-cruise-deepfake-chris-ume-security-washington-dc/100234772.

[2] Neish AS. Mucosal Immunity and the Microbiome. Annals of the American Thoracic Society. 2014;11(Suppl 1):S28-S32. doi:10.1513/AnnalsATS.201306-161MG.