We will live with and alongside intelligent systems, but now is the time to work out the terms of engagement. Tomorrow is too late.

Technology is – again – on the move. The media, both old and new, awash with promises of how AI will deliver disease-free longevity, eyewatering productivity, save our public services, alongside exciting offers ranging from our own personal robot to human settlements on Mars. This tech utopianism is pierced by the occasional outburst from those trying to warn us that AI is powerful, that unfettered it will cost us our jobs, our democracy, our lives – when it takes over.

Amid these conflicting messages is an increasingly distrusting (anxious) public and policymaker paralysis. Paralysis that may be a precursor to the largest transfer of wealth and power from UK plc to a small group of people in the US, and maybe China.

Over the last 2 decades, the internet brought connectivity to two thirds of the world. Developed in academic and military labs, the technology embedded the assumptions of the culture in which it was incubated. The utopianism of the founders was based on the idea that that the tech would be open, there would be no gatekeepers, and all users were equal.

Recent history tells a different tale in which the tech subsumed by business interests, monopolised whole sectors, the public square suffered a series of enclosures, and whole swathes of the population, children, women, ethnic and faith groups found their needs unmet by the assumptions inherent to the design.  

Meanwhile users attached to their devices in an attention economy increasingly dependent on a handful of gatekeepers – or tech oligarchs – who control access to services, content and each other. Monopolising our attention and taking high premiums for giving legacy businesses access to us - through advertising simultaneously taking a large cut from independent traders, app developers or creatives has created an asymmetry of power that has exacerbated rising inequalities and discontent.

Fundamental to AI already widely deployed in the products and services used daily, is data. Data, the least glamourous, most valuable asset in the AI value chain. Data given, inferred, demanded, scraped, hoarded, sold, shared, traded, stolen, is the ingredient on which the power and profit of the tech incumbents is built. And without it, the new generative models would not exist.

Their current businesses and future success of the sectors utopian vision of AI is entirely dependent on harnessing our data. It is estimated that most of the value of AI (estimates range to as much as 90%) will accrue to US and China. In the gold rush, the data rich companies will be the big winners. Almost all the major AI players are now funded or owned by tech giants.

AI is being built – ignoring the very present harms of AI-generated Child Sexual Abuse Material and Mis and Disinformation that at best, makes a distrusting public more distrustful, and at worst, undermines faith in democracy and its institutions. AI available to us all can model the reintroduction the Smallpox virus and will soon allow the creation of new novel viruses.

Meanwhile tasks that could be done by GOFAI (good old-fashioned AI) are now being done by new resource-hungry generative models.  Last year, for the first time, the data centres in Ireland used more energy than all other domestic and industrial demands on the grid.

There is no doubt of AI’s potential to find disease cures, compute large data sets, make radical changes to the process and costs of the state. But there are pressing issues that government needs to grasp. We have extremely valuable data sets that should be characterised as essential infrastructure, we must value the Intellectual Property rights of our creative industries, we should put questions of data processing front and centre of our industrial strategy and we should seek a position that accrues value to UK plc and that considers the vulnerable and protects citizens safety and rights.

There will be accidents. As a member of the US armed forces said to me, “when thinking about the end of the world, I am more concerned about careless releases of powerful systems into the wild than the use of lethal robots”.

Public discourse is dominated by those with conflicts of interest, with little to no interest in setting more stringent terms for AI’s safe development. Tech is too important to leave to the technologists. We will live with and alongside intelligent systems, but now is the time to work out the terms of engagement. Tomorrow is too late.