Elon Musk has been ordered to be interviewed under oath to determine if he made specific statements about Tesla’s Autopilot feature after the carmaker threw their authenticity into doubt, saying the billionaire was often a target of online “deepfakes”.
A judge in California made the tentative ruling after raising concerns that such arguments could be used by Musk – and other high-profile celebrities – to “avoid taking ownership” of their public statements.
The “deepfake” claim was made by lawyers for Musk’s firm while defending a lawsuit brought over the safety of Tesla‘s driver assistance system.
The family of Walter Huang are suing the company at Santa Clara Superior Court over a car crash that killed the Apple engineer in 2018.
Mr Huang’s family argues Tesla’s partially-automated driving software failed.
However, the carmaker contends the engineer was playing a videogame on his phone before the crash and disregarded vehicle warnings.
Lawyers for Mr Huang’s family have pointed to Musk touting the safety of the Autopilot feature – including a 2016 statement in which he allegedly said that Tesla’s Model S and Model X vehicles “can drive autonomously with greater safety than a person”.
Musk’s lawyers say the entrepreneur cannot recall details about statements and have argued that “like many public figures, [he] is the subject of many ‘deepfake’ videos and audio recordings that purport to show him saying and doing things he never actually said or did”.
However, the judge presiding over the lawsuit, Evette Pennypacker, questioned Tesla’s argument, which she described as “deeply troubling”.
“Their position is that because Mr Musk is famous and might be more of a target for deepfakes, his public statements are immune,” Judge Pennypacker wrote.
She added that such arguments would allow him and other famous people “to avoid taking ownership of what they did actually say and do”.
Judge Pennypacker tentatively ordered a limited, three-hour deposition, where Musk could be asked whether he
actually made the statements on the recording.
What is a deepfake?
A deepfake usually involves an image or video in which a person or object is visually or audibly manipulated to say and do something that is fabricated.
They have been used to create fake videos of Barack Obama calling Donald Trump a “complete dips**t” and Facebook boss Mark Zuckerberg bragging about having “complete control” of people’s data.
Deepfakes have also been used to create fake pornographic images of celebrities.
And there are concerns that they could be used to create mass panic or influence elections by creating fake videos of politicians.
Some countries are even attempting to crackdown on deepfakes. In England and Wales, proposals have been put forward to make the creation of pornographic deepfakes without consent a criminal offence.
The images are made using AI technology – though experts are also attempting to use AI to tackle deepfakes by making technology that can more easily pick-out fact from fiction.
California judges often issue tentative rulings, which are almost always finalised with few major changes after such a hearing.
The lawsuit is scheduled to go into trial on 31 July.
It comes after a California state court jury on Friday found Tesla’s Autopilot feature did not fail in what appeared to be the first trial related to a crash involving the partially automated driving software.