You know that you’re in a fabulously expensive store when you find that nothing has a price tag on it. OpenAI and Microsoft could be getting that unnervingly exclusive feeling because they are being sued by the New York Times for unspecified billions. That is all that the lawsuit filed this week in a Manhattan court says: “billions”. How many? That’s unstated, a nod to the tremendous value of artificial intelligence and the intellectual property on which it is trained.
The newspaper is one of several entities suing the owners of ChatGPT for scraping their content without permission to rival their products. The latest interest group is non-fiction writers, led by a case filed by a Hollywood Reporter editor, Julian Sancton; Pulitzer-winning authors have joined in. Suits have also been brought by Jonathan Franzen, George R R Martin and John Grisham. And, of course, the strike in Hollywood had much to do with the existential threat posed by artificial general intelligence.
The NYT case is getting the most attention on account of the volume of its content that has been used to train ChatGPT archives going back over the decades. It is one of the world’s most important records of contemporary history, though the NYT suffers from the same failing as AI hallucinations, or confident assertions about the reality of things that do not exist, like weapons of mass destruction in Iraq, or a dark Chinese plot involving an Indian news portal.
ChatGPT is trained on the internet and its contents, a wide variety of matter on subjects as diverse as PHP and philosophy, wave mechanics and windsurfing. The spread includes all of Western literature, from Sudden to Ulysses, and all of media and the arts. That is where the blowback is coming from. It tells us what is going to be acceptable about AI, and what will be regarded as enemy action.
This was anticipated by Isaac Asimov’s Three Laws of Robotics, which were published in a short story in 1942 and collected in one of the most widely-read anthologies in the history of science fiction, I, Robot (1950). In it, they were attributed to a fictitious handbook of robotics dated to over a century later, 2058. It overshot the AI age by 35 years, but the First Law got it right (the others are derivative): “A robot may not injure a human being or, through inaction, allow a human being to come to harm.”
Generations read that statement to mean that a robot is not allowed to offer violence to humans. Even Asimov may not have anticipated this, but its scope can broaden according to the definition of violence. For instance, it can be read to mean that machines should be disbarred from competing with humans in a way that threatens their essence. And read thus, it becomes rapidly clear that the focus of the discussion on AI has been misplaced.
The economic turmoil that AI threatens by taking up jobs and roles, perhaps within the year, is only that turmoil. Every technological revolution brings on a wave of redundancies. Farm machinery removed peasants from the fields. The washing machine and kitchen appliances conspired to steal the jobs of domestic workers. The computer is a multi-role, multitasking device, so it has been the most devastating so far, taking assistive roles away from humans across industries pool typists, telephone switchboard operators, data tabulators, and accountants. The AI revolution is only sharpening this process.
Also read: The trouble with numbers
Economic turmoil triggered by a technological revolution settled down soon after the event. Some workers move on to where their efforts continue to be rewarded. Some rewind to pre-tech roles, like handicrafts. Technology itself also creates new roles. The few who remain learn to treat the new technology as an assistant. Farmers become combine harvester drivers and seed drill operators. Domestic workers learn to use the microwave and the washing machine. In the short time since ChatGPT was released, its users already have it.
Very often, the first prompt used on the machine is: “Imagine that you are my assistant.” AI is going to find great assistive roles in a variety of fields, from radiology and oceanography to drafting office correspondence and selling retail. This is the real story, but it’s so deathly boring and obvious that the press doesn’t cover it too much.
The anxieties and excitement the strikes and lawsuits begin when AI ventures into intellectual property. First, it is trained on intellectual property owned by creative people. Then it produces its own intellectual property, and human producers of intellectual capital draw the line tight. Other anxieties about unemployment, students cheating and a plague of deepfakes are very real, but apart from issuing grave warnings, no one has acted substantively against them. The first serious challenge to the boom industry of 2024 has come from creative people, who had no trouble understanding that they are being ripped off.
For millenniums, people have been wondering about the essence of humanity. Is it a divine spark? A gift for language and communications, or formally compiling collective memory? The blowback against AI may have given us the definitive answer. The one asset that humans will defend without compromise is creativity. It’s the only one without a price tag. It’s worth the unspecified billions that the New York Times is suing for. If that were farmed out to a machine, human life would be rendered meaningless, and it is only normal for the creative people of the world to close ranks against the machine.
(@pratik_k)
Pratik Kanjilal, Editor of The India Cable