Meltdown May #2: Elon Cashes In His Liar's Dividend
Elon isn't even the founder of the "Deepfake Defense"
One of the most important tools that Elon Musk has leveraged in his rampage across the economic landscape is a total inability to feel shame. From “Full Self-Driving” to “Funding Secured,” Elon’s ability to dig past any semblance of propriety and forward the most absurd and blatantly self-serving positions possible has been central to his success. He may not be the only prominent figure possessed of such abilities (we all have our favorites), but last week he hit a new low that re-establishes him as the most shameless to ever do it.
In the trial over the 2018 death of Walter Huang in an Autopilot-involved crash, Tesla’s lawyers claimed that video evidence of Musk’s deceptive statements about the company’s driving automation technology can’t be relied on because they might be deepfakes. Not that they have evidence that they are in fact deepfakes, just a blanket “maybe they could be, so let’s just dispense with them.” Basically the Shaggy song, only with a 2020s techbro twist.
Now, some readers might be thinking to themselves that this is actually a singularly creative tack, which speaks to Musk’s unique genius. If that’s you, just go ahead and close the tab now and waddle back to the Whole Mars Blog mentions, because things are about to get real bad for you. In fact, Musk came up with the idea the same way he comes up with so many of his most famous ideas: he stole it.
Specifically, he seems to have stolen it from a January 6th defendant, one Guy Reffitt, a “Three Percenter” whose involvement in the storming of the capitol building was captured in what a paper on the topic describes as “overwhelming” audiovisual evidence. Unfazed, Reffitt’s lawyer (and likely future member of Musk’s “hardcore litigation” team) swung for the fences:
“although video, audio, text messages, the testimony of his son, and another insurrectionist clearly an dunmistakably implicated Reffitt in the crimes, his lawyer attempted to undermine the evidence, suggesting the incriminating images may have been fabricated. His lawyer told the jury that the evidence against Reffitt was a “deepfake”–an audiovisual recording created using readily available Artificial Intelligence (AI) technology that allows anyone with a smartphone to believably map one person’s movements and words onto another person.”
This strategy, the paper’s author Rebecca Delfino argues, “is a new danger to our legal system’s… truth-seeking function” to which the existing rules of procedure, ethics and legal precedent offer little defense. The “Deepfake Defense” is part of a broader “Liar’s Dividend” arising from deepfake technology,
“a term coined by Professors Robert Chesney and Danielle Keats Citron to describe the phenomenon that arises when the ability to create convincing fakes allows the image creators to undermine the veracity of real information by claiming that it, too, is a fabrication”
Though relatively new in the legal setting, this tactic has already racked up a track record as sordid as its name suggests: a coup attempt in Gabon, a Republican seeking to discredit the George Floyd video, and the alleged “deepfake cheerleader mom.” There’s even a Rittenhouse Variation of the Deepfake Defense, in which his lawyers (successfully!) objected to iPad video evidence by claiming “this is Apple’s iPad programming creating what it thinks is there, not what necessarily is there.”
In case those celebrity endorsements weren’t enough to convince you, Delfino lays it out in no uncertain terms: “LAWYERS’ USE OF THE DEEPFAKE DEFENSE IS WRONG AND DANGEROUS.” In a brief but fascinating history, Delfino tracks the development of norms against lying layers from the laissez faire “let the Gods deal with liars” attitude of Ancient Greek and Roman courts to the modern regulation of the legal profession by courts, bar associations and state legislatures, concluding that maybe some of this came into being for a reason. Though lawyers are charged to advocate strenuously, “and in some instances… the rules of law and professional ethics give lawyers significant leeway to engage in [deception],” lawyers are also bound to the truth as officers of the court.
From there the paper dives deep into the legal weeds, forcing this non-lawyer to rely on Delfino’s depressingly unsurprising conclusion: “The Procedural Rules and Statutes Are Inadequate.” Once again, it seems, wherever society’s tools of justice are most vulnerable to technological innovation, Elon Musk will eventually show up, his greedy snout snuffling among our institutional rot, extracting whatever he can, whatever the cost.
Luckily, the wonderfully named Judge Pennypacker in Tesla’s case was not impressed with the Deepfake Defense, stating in her ruling [PDF via Ars Technica]:
"Tesla's argument that it cannot commit one way or another to the statements, or in some cases even admit that it is Mr. Musk in the videos, because of the ease with which deep fakes can be made is unconvincing… Mr. Musk was either at these places or he was not; he either said these things or he did not. Ironically, Tesla's refusal to answer these questions only makes a clearer record that Mr. Musk is the only person that has this information to respond to this discovery, one of the pre-requisites to permitting an Apex deposition."
With that, Musk must now appear for a three hour deposition before a plaintiff’s counsel who now knows exactly how little conviction he has in his many, many, many misleading statements about driving automation technology. Needless to say, we imagine said counsel now blithely sharpening their knives with a joyful song in their heart, and we wish them all the best in their endeavors.