AI Needs You is a fascinating read, though not necessarily for the reasons you might expect – what the title suggests isn’t quite what you actually get.
That isn’t to say this is a bad book – quite the contrary, it’s a detailed and informative guide to the history of how the internet was set up, from the earliest days of computing in the 1950s to today’s conundrums.
The historical theme-setting compares the arrival of AI to other transformative innovations such as IVF, the space race – the author applauds ‘the Magna Carta of space’ which disallowed territorial claims to the Earth’s atmosphere – and the nuclear age. The scene-setting is so comprehensive that it takes up most of the book before we arrive in the present. If it had been called ‘AI: How did we get to where we are today?’ then it would get nine out of 10. But it isn’t, it’s called AI Needs You and is sub-titled ‘How We Can Change AI’s Future and Save Our Own’.
Author Verity Harding, director of the AI and geopolitics project at the Bennett Institute for Public Policy at the University of Cambridge, is no ordinary academic. Back in 2013, she was a SPAD (special advisor) to Nick Clegg, the deputy prime minister at the time. Then she left to join Google, where Larry Page was her new boss. She spent a decade at Alphabet, latterly as DeepMind’s first global head of public policy, where in 2017 she co-founded the company’s research and ethics unit, as well as the independent multi-stakeholder organisation, Partnership on AI.
Indeed, since she is closer to AI’s unstoppable momentum than one might at first presume, her book offers many insights into the mindset of a Silicon Valley-enabled writer.
The mores of Silicon Valley are adhered to in AI Needs You – a determination to be positive, with little or no mention of topics that would upset the world’s most sensitive cadre of entrepreneurs. Ergo there’s nothing on why Big Tech should be paying more tax, or why they should be investing in a massive mental health programme to help the tens of millions of workers set to lose their livelihoods to AI, nor any meaningful suggestions as to how governments could establish proper oversight of the AI era – and certainly no mention of the unsustainable power requirements needed to support AI’s deliberations.
Lawmakers must find the will to confront Big Tech because AI screams ‘mortal danger!’ to the very governments that are green-lighting their use (and the democracy they depend on). However, the author does describe the rules which allow the Valley companies to do whatever they like, and acknowledges that if that seemed OK when their schtick began 20 or 30 years ago, today we know enough to realise their business models are not to be trusted.
“No doubt Silicon Valley has a culture problem,” the author writes. “Trust is waning. Greed is winning.” She also describes how Amazon treats its workers – the workers are not being replaced by AI-enabled robots, rather the workforce is being rewired in a different way. “You’re sort of like a robot but in human form,” says one Amazon manager in the excellent introduction to this book. AI needs you indeed!
We’re at a crossroads, says Verity.
“Do we want to set any boundaries at all before we descend into automated, unscrutinised, unaccountable monitoring?” she asks of the current state of play – but the answer is equivocal.
We might want some safeguards for our children and for our jobs, but we’ve no idea how to create them. Verity’s suggestion is to write to elected representatives – MP, council – plus speak to school officials, and join a union to use your voice. All that is perfectly valid but the actualité of AI is that the genie is already out of the bottle, so it’s a rearguard action we’re fighting.
“But,” she writes, “we can begin to set an example for the world by establishing conditions about how AI will not be used. Quick progress could be made, for example, by cracking down on AI-enabled surveillance.” Sounds good, but later we read that “AI-enabled surveillance is perfectly legal in most places, and if there are governments and private clients demanding it, then more of it will get built”. So if in fact there’s zero prospect of a swift ‘crackdown’ on excessive use of AI, the suggestion is perhaps a mere sop for the feeble-minded?
“Without any regulation governing their use,” concludes the author, “educational institutions, private companies and [retail] stores are free to use AI-enabled biometric data programs that can monitor voice patterns, conduct gait analysis, and analyse facial expressions.”
So in what sense can an individual make a difference? At the end of this book, it’s apparent that AI needs you as a source of data – to control, monetise and/or manipulate you. But does the AI industry need or care about your opinions, your insights or your preferences? No, it doesn’t. Are politicians or corporations listening to your concerns? Not really.
The takeaway from AI Needs You (Princeton University Press, £20) is that the challenges of AI won’t be addressed until we understand the scale of its interference in our daily lives already – and by then it could be too late.
