New technology often seems to pop up out of nowhere. You’d never heard of an iPhone, then suddenly you’re shocked when someone doesn’t have one (or bothered by their green bubble). Only a few of your friends have a social media account, then suddenly everybody does. Technology, when it’s good, has a tendency to quickly jump from unique to ubiquitous.
Think of a particular technology that is everywhere today. The hype has pushed it into ubiquity. Everyone recognizes how this technology will change the way we communicate, work, and live. The advancements are whiplash-inducing, everyone you know seems to be playing with it, and the companies building on it seem universal. Capital is focused on the companies building with this technology; meanwhile, the establishment is afraid. Publishing, music, and media are all constantly racing to try and control what feels like a new frontier.
Without any of the specifics, you might assume that this is a description of today’s AI boom. But it’s actually a description of the internet.
From 1990 to 2000 the number of people using the internet exploded from 2 million to 414 million. By the end of the 1990s, the nature of the internet could largely have been defined by one word: open. The ability for anyone to build a website, get fair access to internet traffic, win over users, and provide free expression all represent core pillars of the open internet. But the open internet wasn’t won without a fight.
Over the course of the 1990s, internet service providers (ISPs) like AT&T and Verizon lobbied to be the toll booths of the internet. Disney and Time Warner lobbied for stricter copyright regulations. AOL lobbied for regulations that would benefit its walled garden approach to the internet.
But, by and large, they failed.
Net neutrality made it harder for ISPs to throttle traffic like a mob boss, giving preferential internet speed to the highest bidders. Massive internet protests shut down copyright laws proposed by Disney and Time Warner. AOL was forced to give up its walled garden by sheer user pressure.
There is a near-perfect parallel between the internet revolution of the 1990s and the AI revolution right now. But this time? The establishment is pulling ahead.
This week, with the one-two punch of, first, President Biden’s executive order on AI, and then the AI Safety Summit in the UK, the potential for open and collaborative AI are at risk. People have argued there are legitimate fears with AI, like hallucinations and misinformation, deepfakes for cybercrimes, and the disruption of human jobs. But startups are arguing that the increased friction from these requirements will limit smaller players and further concentrate power in the hands of larger companies like Microsoft, Google, and OpenAI.
And those companies know that. They’ve been the most dogmatic lobbyists in favor of regulating AI. A quintessential example of established companies pulling up the ladder behind themselves. Ben Thompson articulated the dynamic this way:
“If you accept the premise that regulation locks in incumbents, then it sure is notable that the early AI winners seem the most invested in generating alarm in Washington, D.C. about AI. This despite the fact that their concern is apparently not sufficiently high to, you know, stop their work. No, they are the responsible ones, the ones who care enough to call for regulation; all the better if concerns about imagined harms kneecap inevitable competitors.”
While companies like OpenAI and Anthropic sound alarm bells that others have described as AI fear mongering, organizations like Mozilla, Shopify, Hugging Face, and Google Brain founder Andrew Ng have made statements about how we should be worrying about real risks vs. overreacting to the AI doomerism and, as a result, over-indexing to arduous regulations that only large organizations can satisfy.
AI investors and founders from a16z, Benchmark, Meta, Replit, Hugging Face, and more came together to issue a letter to President Biden in response with one key point. “Our primary concern: ensuring AI remains open and competitive, rather than monopolized by a few entities.”
The jury is still out. Will the development of AI be able to follow in the footsteps of the open internet and be able to progress collaboratively around the world? Or will a handful of powerful companies be able to shape regulation in their image and, in so doing, shape the face of AI to reflect their respective values and profit motives? Time will tell.
