Intellectual Property

Faking It: Why Deepfakes Pose Specific Challenges Under Copyright & Privacy Laws

Whether we like it or not, deepfakes are here to stay, and will need to be handled in more solid ways than through the existing copyright framework or patchwork of state laws.

By  
on

(Image via iStock)

You’ve probably seen them by now, and if you haven’t, you will probably ask yourself the same question everyone else does, which is:  how do they do it?  For the uninitiated, I am talking about “deepfakes” (an interesting combination of the phrase “deep learning” and the word “fake”).  Described as “a technique for human image synthesis based upon artificial intelligence,” deepfakes are altered photos or videos that are definitely not what they seem, and in the context of copyright law, disrupting more than you think.

First, some understanding of deepfakes would be helpful.  Altering images and video to be something other than originally intended is nothing new, but these altered images and videos are not your garden-variety staple — these are (usually) created by using a machine learning construct known as generative adversarial network.  Essentially, it involves two artificial neural networks working off each other based on a specific training objective (say, the creation of an image indistinguishable from the original), with the generative network creating new images compared by the discriminative network until the generative network meets the objective.  Of course, that is a gross simplification, but you get the gist — deepfakes are very convincing computer-generated “fakes.”  Don’t believe me?  Look at an example here.  Make no mistake — although there is still some room for improvement, there will likely come a time when such deepfakes will be extremely hard, if not impossible, to detect.

Such convincing fakes are an amazing feat of technology, but present a host of thorny issues when it comes to their use. Although an interesting means for parody, the Mark Zuckerberg deepfake demonstrates how this technology can be used to create a false narrative attributable to a person who never said it.  It can also be used to morph existing images into ones that may depict the original content (or the copyright owner) in manner that is disparaging if not extremely damaging to the work itself.  In a nod to conspiracy theorists, it may not be outside the realm of possibility in the near future that such deepfakes could be used to induce anything from stock market instability to all-out war.  It’s not hard to imagine bad actors using this technology to further their own objectives, and unfortunately, the law has a lot of catching up to do.

The issues presented by deepfakes under copyright and privacy laws are troubling, and here are a few examples that should invoke particular concern:

  • Fair Use.  Although the owners of copyrighted works enjoy certain exclusive rights to those works, the fair use doctrine allows for freedom of expression by permitting the unlicensed use of copyrighted works by others under limited circumstances, such as for criticism, comment, news reporting, teaching, scholarship, and research.  Section 107 of the Copyright Act outlines these criteria, as well as four factors for consideration in the fair use analysis. It should come on little surprise that many deepfakes may fit squarely within enumerated exceptions, but when it comes to weighing the four factors, whether the deepfake is sufficiently transformative or one whether the use would have a negative effect on the market for the work are far more complicated elements to consider. Needless to say, deepfakes can take the fair use analysis to a whole new level, much to the chagrin of the copyright owner.
  • Digital Millennium Copyright Act (DMCA).  Enacted in 1998, the DMCA provides (among other things) a mechanism for copyright owners to request a “takedown” of their copyrighted content from websites, but this mechanism is not perfect.  For example, the fair use doctrine listed above would be a valid defense to such a takedown request. Further, such DMCA take down notices are only as valid as the country that recognizes them, so they are not much help if the website hoisting the infringing work is based in a far-off jurisdiction that does not recognize them.  Moreover, it is not a stretch to assume that many deepfakes will be posted through social media channels such as YouTube and Facebook to name a few.  Given the recent postings of deepfakes on social media platforms already, there does not seem to be any universally accepted way of handling such deepfakes, so each social media platform will handle it differently.

Sponsored

  • State Privacy Torts.  Most states either recognize at common law or have enacted laws that address violations of individual privacy, generally addressing (i) intrusion upon seclusion or solitude, or into private affairs; (ii) public disclosure of embarrassing private facts; (iii) publicity which places a person in a “false light” in the public eye; and (iv) misappropriation of one’s name or likeness for commercial gain.  To the extent a deepfake crosses the threshold of liability under an applicable state tort, the individual damaged by the deepfake may seek redress.  That said, not all states recognize all the aforementioned privacy torts.  Moreover, redress may rest with the individual(s) depicted in the work as opposed to the copyright owner.

Whether we like it or not, deepfakes are here to stay, and are only going to get better and more convincing. As a result, the issues presented will need to be handled in more solid ways than through the existing copyright framework or patchwork of state laws.  Only time will tell if this will be accomplished through updates to the Copyright Act or through other federal legislative measures.  Let’s just hope it is sooner rather than later — until then, it seems that the law just may be faking it itself.

Tom Kulik is an Intellectual Property & Information Technology Partner at the Dallas-based law firm of Scheef & Stone, LLP. In private practice for over 20 years, Tom is a sought-after technology lawyer who uses his industry experience as a former computer systems engineer to creatively counsel and help his clients navigate the complexities of law and technology in their business. News outlets reach out to Tom for his insight, and he has been quoted by national media organizations. Get in touch with Tom on Twitter (@LegalIntangibls) or Facebook (www.facebook.com/technologylawyer), or contact him directly at tom.kulik@solidcounsel.com.

Sponsored

Topics

, , ,