200 years ago this month, a 17-year-old Mary Shelley gave birth to a premature daughter, who died at the age of just two weeks. The tragedy was deeply harrowing for Shelley, but it was also a vital part of the history of English literature. Three years later, Shelley published Frankenstein; or, The Modern Prometheus — the book that created not only the indelible title character and his monster, but also the entire genre of science fiction, and the idea that great and wonderful advances in science can have terrible consequences.
Shelley’s idea that modern technology can have a dystopian downside was a powerful one, which never died. Indeed, after two World Wars, one of which was ended by use of the atomic bomb, it became conventional wisdom.
Today, nearly all major technological advances are accompanied by fear. As Douglas Adams put it, “Anything that is in the world when you’re born is normal and ordinary and is just a natural part of the way the world works. Anything that’s invented between when you’re 15 and 35 is new and exciting and revolutionary and you can probably get a career in it. Anything invented after you’re 35 is against the natural order of things.”
Genetically modified food, for instance, is harmless — yet only 37% of US adults consider it safe to eat. And while you might think that many of the big technology successes in recent years have been relatively unthreatening (as Peter Thiel says, “we wanted flying cars; instead we got 140 characters”), that hasn’t stopped Black Mirror from becoming a powerful cultural meme. Now, a much more powerful wave of new technology is fast approaching.
Consider a recent long interview with Demis Hassabis, for instance, the founder of artificial intelligence company DeepMind. Barely has the interview even begun before the journalist brings up the fear that smart computers could mean the end of human life:
The uncertainty lies in whether these artificially intelligent beings will be motivated by a desire to guide and assist us or simply to do away with people like old gadgets that have served their purpose. Also on the YouTube video, Hassabis describes his AI computer playing a boxing game in which, after a few seconds of sparring, it corners the opponent and pummels them into submission. The audience laughs as Hassabis explains that the computer “ruthlessly exploits the weakness in the system it has found”. But perhaps this is an apt analogy. As physicist Stephen Hawking wrote last year, AI would be “the biggest event in human history . . . unfortunately, it might also be the last.”
Hassabis, sensibly, dismisses such concerns out of hand. Other science, however, has managed to frighten even its inventors. Jennifer Doudna, for instance, is at the forefront of one of the most exciting biomedical advances in living memory: engineering the genomes not of plants, but of people. Her cheap and easy Crispr technology holds out the promise that anybody with a gene defect could get that problem fixed, on an individual, bespoke basis. No more one-size-fits all disease cures: everything can now be personalized. The dystopian potential here, of course, is obvious: while Doudna’s name isn’t Frankenstein, you can be sure that if and when her science gains widespread adoption, the parallels will be hammered home ad nauseam.
Doudna is particularly interesting because she doesn’t dismiss fearmongers as anti-science trolls. While she has a certain amount of control over what her own labs do, her scientific breakthrough is in the public domain, now, and already more than 700 papers have been published in the past two years on various aspects of genome engineering. In one high-profile example, a team of researchers found a way of using Doudna’s breakthrough to efficiently and predictably cause lung cancer in mice.
Because Doudna’s technology allows anybody with basic molecular biology training to play around with engineering a human genome, it would be surprising if, somewhere, such experiments didn’t rapidly advance to human experimentation. Or even just scientists experimenting on themselves, given that the cost of sequencing an individual’s genome is now just $1,000.
Meanwhile, at the more mundane end of the technological spectrum, technology companies like Uber and Airbnb have become multi-billion-dollar behemoths in large part through a strategy of aggressive regulatory arbitrage. As venture capitalist Balaji Srinivasan says, one huge reason for Uber’s success is that its founders, faced with orders that it immediately cease operations or face up to 90 days in jail per day that the company kept on operating, made the decision to simply ignore those orders. We all know how that story ended: by flouting the law, not only in San Francisco but in dozens of other cities around the world, Uber’s founders became billionaires.
Or look at Airbnb, which has been fighting regulators in New York City since 2010. Most of the Airbnb listings in New York are illegal, and yet New York remains Airbnb’s second-biggest market. Given the choice between having thousands of hosts breaking the law, or not having those hosts at all, Airbnb is clearly opting for the former. The only words of caution the company gives to hosts are buried two clicks deep on a page about “responsible hosting”.
What all of these cases have in common is that the existing regulatory infrastructure simply can’t cope with a lot of modern technology. If artificial intelligence did pose a real threat to society, there’s no world technology regulator who could inhibit it at all. When I asked Doudna who, if anybody, was in charge of regulating genome engineering, she drew a blank — even answering that question in the U.S. is non-trivial, and globally it’s impossible. (Already it’s China, rather than the U.S., that leads the world in gene sequencing technology.)
We’re long overdue for a regulatory revolution. It won’t come from Silicon Valley, home to techno-libertarians like Thiel, who see all regulation as unwarranted meddling in private affairs. It might well, on the other hand, be born in Europe, which has long since realized that a lot of regulation belongs one level up from the nation-state. If there is to be any effective regulation of global technologies, from genome engineering to Google, it’s going to have to take place at a supra-national level. Indeed, what we really need, to borrow a concept from Silicon Valley, is a global regulatory platform.
Such a platform would have the happy effect of lowering the barriers to entry in many technological fields. At the moment, there is a relatively small number of deep-pocketed and aggressive companies like Uber, who have both the ability and the belligerence to go head-to-head with regulators in hundreds of cities around the world. The result is that they crowd out potential competitors, giving themselves a potentially unassailable first-mover advantage. A similar syndrome can be seen in a lot of financial-services companies, from PayPal to Stripe: the number of regulators you need to deal with is enormous, even just within the U.S., and very few small companies have those kind of resources.
But what if every city, state and country were able to compile its regulations on a global regulatory platform? At that point, a company which complied with one set of regulations would automatically be deemed to be in compliance with identical regulations in all other jurisdictions. And a huge part of any company’s regulatory headaches would disappear.
My hope for the next few years, then, is that we see a regulatory revolution commensurate with the technological revolution that is going to happen anyway. There are lots of governments and organizations which should at least be talking about such a thing, including the UN, EU, US, G-20, and many others. The idea would not in the first instance be to build a massive new global bureaucracy, but rather just to build an extensible platform which existing regulators could plug into.
In this new world, sharing-economy companies like Uber and Airbnb could be regulated sensibly, in a way that recognized their global clout without trying to make them fit into 20th-Century structures. Both companies, I’m sure, would be ecstatic to be able to work within a single global regulatory regime. They could agree to do things like collect local taxes, comply with privacy regulations, and have clearly-defined avenues for redress when things go wrong. Uber could implement externally-defined vetting requirements for its drivers, and agree to rules surrounding the registration and ownership of the cars they drive; Airbnb could agree to a maximum number of homes that could be rented out by any one individual.
To put it another way: the parts of today’s taxi and hotel regulations that sensibly apply to Uber and Airbnb could be applied; the parts that don’t really make sense in a sharing-economy world could be easily dropped; and new regulations around things like big data, which do make sense for platforms but which don’t make sense for smaller businesses, could easily be added.
More importantly, a flexible global regulatory platform could be the only way to be able to keep some kind of handle on the ultra-distributed science of things like genome engineering. Nothing could prevent a malicious actor secretly opting out of the system altogether, of course. But even a simple protocol which would require all scientists to say what they are working on — and, crucially, to share their negative results — would be a vast improvement over what we have right now, which is nothing.
Such a protocol might also increase public trust in science. When the public looks at close-mouthed multi-billion-dollar for-profit companies like Monsanto, it’s easy for them to take what the company’s scientists are saying with a grain of salt. But at the moment we have little choice but to trust the private companies, since public regulation is so woefully weak. A global regulator with international funding and unimpeachable expertise would not be trusted by everybody, but it would at least be able to define a legal and ethical consensus — and its authority would make it much easier to rein in any scientists acting in a dangerously reckless manner.
Something along these lines has already happened in bank regulation: national central banks all speak the language of the Basel international banking regulations, which have real teeth. So let’s hope that we see a similar development in many other areas, too. It might not prevent Frankenstein’s monster from ever being built. But it should at least reassure the public that we’ll have ample warning if anybody starts moving in that direction.