An AI Future Is Much Shakier Than You Think

We are often told these days that the world stands at a technological crossroads; that we are hurtling forward into a future shaped by artificial intelligence tools such as ChatGPT. I suspect, instead, that 2024 will be the year we are reminded of the Ghost of Napster—and other failed digital futures.

If you’re under 35 or over 60, you probably don’t know what Napster was. But for a moment, it was the future. Napster was a peer-to-peer file sharing service. For a few years—roughly 1999 to 2002—it had an aura of futurity about it. The basic logic of Napster was simple enough: Somebody buys a CD. They upload the songs to their computer. The computer is connected to Napster’s peer network. Napster renders those songs freely available for download to every other user of the network. You could think of this a bit like making infinite mix tapes and distributing them everywhere. But, from the perspective of the music industry, it was also quite a bit like theft.

The music industry trembled, and for good reason. What would happen to music sales as the cost of digital reproduction and distribution fell to zero? If all music everywhere is free, then how are professional musicians supposed to make a living? What started with music could easily spread to film and television. The creative industries were supposed to be protected by copyright law. But what did 20th century copyright law even mean for 21st century communications technologies?

The near-consensus answer back then was that these new communication technologies were inevitable. If there was a conflict between old copyright law and new media, then copyright would have to bend as a result. Electronic Frontier Foundation founding member and Grateful Dead lyricist John Perry Barlow summed it up in an October 2000 article: “Practically every traditional pundit who’s commented on the Napster case has, at some point, furrowed a telegenic brow and asked, ‘Is the genie out of the bottle?’ A better question would be, ‘Is there a bottle?’ No, there isn’t.”

The Recording Industry Association of America (RIAA) disagreed. It launched countless lawsuits—against Napster, against Napster clones, even against individual users. But those lawsuits seemed like the last gasps of a technologically irrelevant industry. To borrow from Gramsci, the old music industry was dying, while the new struggled to be born. The future would surely be filled with limitless, identical copies of copyrighted works, shared for free among everyone.

The lawsuits did not bring an outright end to file sharing, but they did alter the trajectory of the digital future. The RIAA came across as villains in the public eye, but it also succeeded in adding an aura of uncertainty to pirated music. It couldn’t shut down every file sharing site, but it made it effectively impossible for any individual file sharing hub to develop a profitable business model.

And then, while those lawsuits were still ongoing, Steve Jobs walked into a room with the heads of every major record label and told them, basically, that they were taking one dollar per song, Apple was going to get a cut, and that this was better than nothing. The recording industry agreed, and music in the ‘00s went on to be defined by the iPod. Then the iPhone took the iPod’s place, and an iPhone app called Spotify found a way to monetize music streaming.

The world we inhabit today barely resembles the one that seemed inevitable during the Napster years. Copyright law did not bend to accommodate new technologies. The industries built upon those new technologies bent to accommodate copyright law instead. The new status quo hasn’t been great for musicians or artists. Their interests weren’t represented at the bargaining table, and it shows. Musicians in the ‘00s got pennies on the dollar from iTunes sales—more than the zero they received when music was pirated, but hardly enough to live on. Today’s struggling artists make a fraction of a cent from Spotify streams.

Consider the parallels to generative artificial intelligence (AI) today: Much like Napster, ChatGPT has spread fast, practically overnight, seemingly fueled by college students. Like peer-to-peer file sharing, these AI models exist in a legal gray area. They are trained on massive corpuses of copyrighted works and promise to produce competing works without compensating existing copyright-holders.

The lawsuits have already been filed. At issue is whether these AI systems should be protected under the doctrine of fair use. This is a question that a great many lawyers are spending a substantial amount of time working out. I am not a lawyer and will not pretend to know what the doctrinal answer ought to be. These are not simple legal matters.

But I am reminded of the old aphorism, attributed to Mark Twain: “History doesn’t repeat itself, but it does rhyme.”

The story that I often hear from AI evangelists is that technologies such as ChatGPT are here, and they are inevitable. You can’t put this genie back in the bottle. If outdated copyright laws are at odds with the scraping behavior of large language models, then our copyright law will surely need to bend as a result.

And to them I can only say: Remember the Ghost of Napster. We do not live in the future that seemed certain during the Napster era. We need not live in the future that seems certain to AI evangelists today. Right now, ChatGPT itself is a money-losing proposition, one that racks up losses with every question. It’s possible that it will be another Amazon, turning initial losses into monopoly power. But it might also be another WeWork, a company that so heavily inflated its own revenue projections that it couldn’t break even in today’s rental market.

Keep in mind it was only a year or two ago that the inevitable future of music and art was supposed to be Web3 and the mighty blockchain. NFTs weren’t just going to be receipts for pictures of cartoon apes, they were also going to be a new payment mechanism that funded artists. The whole thing evaporated once the crypto bubble burst. Digital futures are flimsy things.

The Ghost of Napster whispers that the trajectory of no technology is inevitable. New technologies are not exempt from old laws. Some digital disruptors can spend years skirting existing regulations (see, for instance, Uber and the entire gig economy), but copyright catches up faster than labor regulations. The trajectory of any emerging technology is not inevitable, especially when its intended trajectory undermines the interests of existing industries. Copyright law doesn’t bend to accommodate your vision of the digital future—the digital future bends to accommodate copyright law.

The boundaries of our creative industries are being renegotiated. Let’s hope that, this time, the artists themselves have a seat at the table