Policymakers need a plan to contain AI or risk dire consequences
4 min read
One way of thinking about the last 500 years is as a steady spreading of power. Thanks to a slowly accreting set of political, social, economic and technological measures, power dispersed out from its traditional centres.
Over time, more people had more capacity to do things. They had more wealth and more rights, had access to and ownership of tools and capabilities that for generations before were unthinkable. From the printing press to the right to vote, dishwashers to the minimum wage, the tight clustering of power in a limited number of people and institutions gradually but remorselessly eroded. However you understand it, power changed and diffused.
This is a revolution so profound that it threatens the foundations of the nation state itself
This process will, I think, continue. But its slow and linear character is about to decisively shift. The next couple of decades will see the fastest and most dramatic plummeting in the “cost” of power we have ever seen. This is a revolution so profound that it threatens the foundations of the nation state itself. It will be a phase of transition in this long historical process, empowering and enabling everyone to achieve their goals, whatever they are. The kind of services and capabilities currently reserved for national governments and billionaire CEOs will be available to everyone, from retirees in Derby to street kids in Delhi. Countless day-to-day activities from organising a birthday party to crafting an effective business strategy will become much easier. But bad actors will also have access to these tools, with dire consequences.
In my book, The Coming Wave: AI, Power and the 21st Century’s Greatest Dilemma, I argue that technologies like AI and synthetic biology will be behind this change. Only now emerging as forces to be reckoned with (as opposed to niche research interests), they will reach full maturity over the next decade or so. Imagine, for example, an AI that doesn’t just say something – it can do it, it can achieve goals autonomously. This is what I call “artificial capable intelligence” and will be a hugely significant milestone in the journey of AI. You will be able to set some arbitrary goal and, within certain parameters, your AI can get it done. Want to make £1m dropshipping on Amazon? Leave it to the AI. In this world, everyone will have an A-team around them – the best lawyers, doctors, coaches, strategists. Everyone has world-class backing.
If computers and the internet were largely about information and communication, this wave of technology is about making things happen in the real world. Whether it’s autonomous cars or gene-edited crops or increasingly ubiquitous clean tech delivering abundant energy, it is an enabler of actions on an epic scale. It’s this aspect of the coming wave that makes the near future such a time of immense opportunity and peril.
Technology will deliver a colossal boost in wealth even as it automates jobs; unleash creativity as well as a torrent of misinformation; create new business as well as opportunities for destructive assaults. AI will plan healthcare systems and invasions, design cures and diseases alike. Whatever force you want amplified, it can – in theory – amplify it. No wonder this could be the most productive time in human history. However it could also lead to a dark series of trade-offs between extreme surveillance and catastrophe.
As we head into the UK’s AI Safety Summit there will be many issues on the table. They are all significant. But it’s important not to lose sight of this attribute of dispersing power, not just how it forms part of a much older societal journey, but how it supercharges it. This unacknowledged feature explains so much of the fear around AI that the summit will try to address. If it can take steps towards what I call containment, where society keeps a meaningful grip on the technologies it produces, can hold the benefits of this dispersal while eliminating the risks, then it will be a success. That is a goal deserving all our support.
Mustafa Suleyman, co-founder and CEO of Inflection AI, former vice-president of AI products and policy at Google, co-founder of DeepMind
PoliticsHome Newsletters
Get the inside track on what MPs and Peers are talking about. Sign up to The House's morning email for the latest insight and reaction from Parliamentarians, policy-makers and organisations.
