TheChek News Flow

Legal scholars, lawmakers and at least one Supreme Court justice agree that companies will be liable for the things their AIs say and do—and that the lawsuits are just beginning.

If your company uses AI to produce content, make decisions, or influence the lives of others, it’s likely you will be liable for whatever it does—especially when it makes a mistake.

This also applies to big tech companies rolling out chat-based AIs to the public, including Google and Microsoft, as well as well-funded startups like Anthropic and OpenAI.

“If in the coming years we wind up using AI the way most commentators expect, by leaning on it to outsource a lot of our content and judgment calls, I don’t think companies will be able to escape some form of liability,” says Jane Bambauer, a law professor at the University of Florida who has written about these issues.

The implications of this are momentous. Every company that uses generative AI could be responsible under laws that govern liability for harmful speech, and laws governing liability for defective products—since today’s AIs are both creators of speech and products. Some legal experts say this may create a flood of lawsuits for companies of all sizes.

It is already clear that the consequences of artificial intelligence output may go well beyond a threat to companies’ reputations. Concerns about future liability also help explain why companies are manipulating their systems behind the scenes to avoid problematic outputs—for example, when Google’s Gemini came across as too “woke.” It also may be a driver of the industry’s efforts to reduce “hallucinations,” the term for when generative AIs make stuff up.

The legal logic is straightforward. Section 230 of the Communications Decency Act of 1996 has long protected internet platforms from being held liable for the things we say on them. (In short, if you say something defamatory about your neighbor on Facebook, they can sue you, but not Meta.) This law was foundational to the development of the early internet and is, arguably, one reason that many of today’s biggest tech companies grew in the U.S., and not elsewhere.

But Section 230 doesn’t cover speech that a company’s AI generates, says Graham Ryan, a litigator at Jones Walker who will soon be publishing a paper in the Harvard Journal of Law and Technology on the topic. “Generative AI is the wild west when it comes to legal risk for internet technology companies, unlike any other time in the history of the internet since its inception,” he adds.

I spoke with several legal experts across the ideological spectrum, and none expect that Section 230 will protect companies from lawsuits over the outputs of generative AI, which now include not just text but also images, music and video.

And the list of potential defendants is far broader than a handful of big tech companies. Companies that simply use generative AI—say, by using OpenAI’s tech as part of a service delivered to a customer—are likely to be responsible for the outputs of such systems. This is a potentially vast universe of services, ranging from evaluating investments to providing customer support.

Among the most compelling indications that companies won’t be protected by current law is Supreme Court Justice Neil Gorsuch’s early 2023 statement on the subject.

In discussing a Section 230 case before the court, he said: “Artificial intelligence generates poetry. It generates polemics today that would be content that goes beyond picking, choosing, analyzing or digesting content. And that is not protected.”

And as companies like OpenAI argue in legal briefs over whether scraping copyrighted content from the internet counts as theft of intellectual property, they may actually be hurting their case that they aren’t responsible for the content their systems produce.

Some AI companies have argued that their AIs “substantially transform” all the content they are trained on. That means, they argue, that they don’t violate copyright protections, under the doctrine of fair use. If that is true, it would seem to indicate they are “substantial co-creators” of the content they are displaying. That is the point at which a company is no longer merely hosting content, and loses the protection of Section 230, says Ryan.

Normally, when companies perceive a gap in existing laws, they lobby Congress for a fix. But lately, Congress has been keen to strip away some of the protections Section 230 already offers, by specifying that companies can only have Section 230 protection if they play by certain rules. Congress’s current mood is the opposite of what companies that make and use AI want.

“In the discussion about, ‘What do you want to do about Section 230’—the answer that many lawmakers will give you is ‘We should gut it,’ ” says Adam Thierer, a senior fellow at the R Street Institute, a conservative think tank. As a result, companies are eager to simply hold on to the existing law and not agitate for any changes, he adds.

The complicated work of figuring out how to apply old laws to AI might be best accomplished case-by-case in courthouses, says Bambauer of the University of Florida. “I’m even going to go out on a limb and say maybe it should,” she adds.

Her argument is, essentially, that by letting cases play out as new information about the harms and benefits of AI is revealed, companies might have the freedom to innovate—but could be reined in when they go too far.

OpenAI is being sued for defamation in at least two cases, including one in which a Georgia radio host alleges that the company’s chatbot wrote an answer that falsely accused him of embezzlement. The company has argued that it isn’t responsible for what its chatbot creates, because its product is more like a word processor, in that it is a tool people use to create content.

In a filing, OpenAI responded to the suit. “By its very nature, AI-generated content is probabilistic and not always factual, and there is near universal consensus that responsible use of AI includes fact-checking prompted outputs before using or sharing them,” its lawyers wrote.

The argument that ChatGPT is more like a word processor than an actual creator of content is likely to fail, says Jason Schultz, director of New York University’s Technology Law & Policy Clinic. “Microsoft Word gives you a blank document, it doesn’t give you a pre-scripted essay,” he adds.

Speech is one kind of liability for companies using generative AI. The design of these systems can create other kinds of harms—by, say, introducing bias in hiring, giving bad advice, or simply making up information that might lead to financial damages for a person who trusts these systems.

Because AIs can be used in so many ways, in so many industries, it may take time to understand what their harms are in a variety of contexts, and how best to regulate them, says Schultz.

Meanwhile, the legal uncertainty for companies that use generative AI, and resulting compliance issues and litigation, may create an unsustainable legal risk for many, says Ryan of Jones Walker. Some, like Therier of R Street Institute, believe this is a threat to the development of the entire field of AI.

Others think that making the legal threat manageable may be as simple as limiting how today’s generative AI tools are used. If making AI chatbots and things that resemble them leads to too many lawsuits, the companies developing the underlying AI technology may simply cut off access to it, says Michael Karanicolas, executive director of the Institute for Technology, Law & Policy at UCLA.

“If we have these tools, and large volumes of people are doing dangerous things as a result of receiving garbage information from them, I’d argue it isn’t necessarily a bad thing to assign cost or liability as a result of these harms, or to make it unprofitable to offer these technologies,” he adds.