AI: a bold new frontier or the end of civilisation?

 In my July briefing for the Jimmy Reid Foundation, I examined the planned growth of AI data centres in Scotland. I didn't have space to develop the section on the merits or otherwise of AI, and many of those opposing AI centres accept that AI is here to stay but oppose hyperscale data centres. As the famous environmental campaigner (and now AI centre opponent) Erin Brockovich says, “That genie is out of the bottle: it’s here, it’s an effective tool, you can use it or not.”

Street art in Tel Aviv (Artem.G, CC BY-SA 4.0)

Personally, I am sceptical that you can separate the two issues. If you accept that AI is here to stay, can you reasonably oppose the means to deliver the benefits? For example, those popular image-generation apps are much more energy-intensive. That is not to downplay the justified concerns of residents facing these massive developments on their doorstep. Concerns that I doubt will be assuaged by the Scottish Government championing 'green' data centres. Some accept that AI cannot be 'put back in the bottle' but argue for greater regulation and hope that technology will provide an answer to the environmental impact, such as faster microchips. This sounds like wishful thinking to me, an argument we heard before over climate change, while the planet is burning. 

The EU has introduced some legislation to regulate AI. The UK does not have any AI-specific legislation. Instead, AI is regulated in the context of its use through existing legal frameworks, such as financial services legislation. It is argued that if democracies limit AI, autocracies like China will win the arms race. I am sure the naval chiefs in the Dreadnought era will be familiar with this argument!

Supporters argue AI can improve efficiency, automate repetitive tasks, solve complex problems, and support advances in healthcare, education, science, and business. It can help people make better decisions by analysing large amounts of data quickly and can perform tasks that are dangerous or difficult for humans. Critics argue that AI may replace some jobs, increase inequality, and pose risks to privacy, security, and the spread of misinformation. There are also concerns that AI systems can make biased decisions, be misused, or operate without enough human control and accountability. Not to mention destroying books!

There is an interesting interview in the latest RSA journal with Cory Doctorow, the technology activist, who argues that the unit economics of AI are really poor. "Every generation of AI is more expensive and money-losing. And they don't have a story about how the economics improve." He warns that when AI crashes, there will be lots of politically angry people, paving the way for authoritarian governments. My immediate thought was that this is just the outcome someone like Palantir's Peter Thiel is planning for. Doctorow argues that the 'regulation will ensure China will win' thesis is simply a cover for the monopolist behaviour we already see in companies like Apple. Stephen Boyd also makes a case for AI regulation in The Herald.

Another perspective on AI comes from a source that few would regard as a liberal leftist. Carson Block is a short seller and the chief executive of Muddy Waters Research in the USA. Writing in The Economist, he argues that in a few years, artificial intelligence will displace a significant portion of the world’s highly paid knowledge workers. These workers will withdraw funds from their investments, ironically causing a big price drop in AI firms. I would also say look out for pension funds that are heavily invested in these firms through increasing passive investment strategies. I had thought that the 'Jevons paradox' would save us (new technology that increases efficiency also increases demand, thereby creating new jobs that offset losses), and scepticism about the slope of the adoption curve. Block says both will be proven wrong, leading to a massive financial crisis. He concludes, "We will witness not just an almighty market correction, but the end of the existing social contract."

Surveys indicate that most Americans, who are at the cutting edge of this change, have a similarly gloomy view. A YouGov poll found that a whopping 45% say they are very or somewhat concerned that AI might cause human extinction. This is driving a rare cross-party political consensus that AI is a problem, not just a driver of growth. However, solutions are in shorter supply. Seventy-nine municipalities in the US have so far issued moratoriums on new data centres. The counterargument is that this backlash is dangerous. AI could power a surge in productivity and incomes, help find cures for untreatable diseases and improve everything else. 

The speculation over AI is justified and interesting, but what is needed is better facts. Investigate the data on energy and water use, as well as job security. Regulate where practicable, building international cooperation where possible, as Yvette Cooper argues this morning. If green data centres are essential, win communities over by making them as 'green' as possible, and offer them a share of the funding. Hopefully, a better system than most renewable generation. While I am not yet convinced that a basic income will work, we should at least be testing ways to share prosperity and tax wealth more effectively. 

And most of all, challenge the inevitability of corporate power by reclaiming collective power to challenge the machine. The Luddites were misunderstood. They rebelled against the way elites were deploying new technology to destroy livelihoods and lower living standards. As should we.