Skepticism in the age of AI deepfakes erodes defamation claims

Artificial intelligence technology is advancing rapidly and has few precedents these days. Among its many uses (and misuses) are generated images, video, and synthesized speech audio, which are increasingly indistinguishable from the real thing. This technology will become more practical and accessible in the coming years.

AI technology is said to make it easier to lie by allowing fake images to be easily disguised as real. But there is also a flip side. It will become more and more reasonable not to believe the veracity of real images and other digital evidence.

Videos and images that today look like irrefutable evidence may soon become of little importance by the public accustomed to how similar evidence can be produced in seconds with a simple text prompt.

This decline in the reliability of digital evidence will have many ramifications for our legal system. Solid evidence was so easily fabricated in the past that courts may now place more emphasis on the credibility of those who produced or authenticated the evidence.

And given the historic role that prejudice has played in jurors’ assessment of the credibility of witnesses, we must be careful not to back down when there is still much work left to do.

Impact on defamation law

Declining reliability of digital evidence poses a particularly thorny issue in defamation litigation. As it becomes harder for people to assess for themselves whether a digital image is truly real, the media will play an increasingly important role in helping the public understand what is real. But maintaining credibility in the media requires accountability, and libel laws play a key role in this.

In most defamation cases, prominent plaintiffs must meet the standard of “actual bad faith” by subjectively proving the defendant Knew the defamatory statement was false or recklessly ignored the truth. It is not enough to show that a defamation defendant should. That is, the plaintiff must, at a minimum, prove by clear and convincing evidence that the defendant was in fact aware of the possible falseness of his statement.

As Justice Neil Gorsuch observed, this was a very difficult burden, and that “over time, the standard of actual malice has reverted from a high hurdle to an effective immunity.” and provoked strong condemnation from media lawyers.

The rise of AI-generated content may make it even more difficult to meet the criteria for actual malice. A defendant’s subjective intent may be evidenced by circumstantial evidence, such as showing that the defendant considered conflicting evidence before publication. In these circumstances, the jury can infer that the defendant knew the defamatory publication was false.

But as the credibility of hard evidence declines, it becomes more difficult to justify this inference, especially under the applicable “clear and compelling evidence” standard. In a world where anyone with a phone can create fake videos, pictures, or audio recordings, isn’t it reasonable, or at least reckless, for a defendant to genuinely question the veracity of previously irrefutable evidence? Isn’t it?

A similar problem can arise when a defendant proves actual malice by showing that he has deliberately avoided the truth. In his 1989 judgment of the U.S. Supreme Court, Harte-Hanks Communications v. Connaughton, the defendant published a defamatory article alleging that the plaintiff offered a bribe in the conversation. Plaintiff provided Defendant with a tape of the conversation to rebut his allegations, but Defendant did not hear the tape.

The Supreme Court dismissed defendants’ excuse that sources said the tapes were not worth listening to because plaintiffs selectively started and stopped recordings, and such manipulation was evident simply by listening to the tapes. However, as sophisticated tools for mimicking voices began to emerge, failure to review evidence when sources said it was fake could result in future courts Blaming reporters can be difficult.

In this way, even if AI technology makes it possible to rely on news organizations to authenticate images and recordings, it would be very difficult to prove actual malice and appear unaccountable. Technology can undermine public confidence in the news media.

Future impact

There has been a growing call in recent years to overturn the actual bad faith standard, fueling claims that it largely exonerates the media from liability. Indeed, suing the media is difficult, but not impossible.

But if AI-generated content undermines plaintiffs’ ability to prove actual bad faith, even in strong cases, criticism of the actual bad faith standard could gain even more momentum. The erosion of actual malice standards caused by AI-generated content is unlikely to be confined to the context of AI.

what to do? On the technical side, efforts such as the Content Authenticity Initiative seek to restore the authenticity of digital content by developing standards for authenticating image and video files. Legally, it’s too early to say how the courts will adapt to this changing landscape, but the challenge of adapting the law to new technology is nothing new.

Courts have adapted other technologies such as fingerprints, DNA, polygraphs and even photography. At this early stage, the attorney must maintain complete documentation to prove the evidence to forestall new suspicions.

This article does not necessarily reflect the opinion of Bloomberg Industry Group, Inc., publishers of Bloomberg Law and Bloomberg Tax, or their owners.