Save articles for later
Add articles to your saved list and come back to them any time.
Man-Made: How the Bias of the Past is Being Built into the Future
Simon & Schuster, $34.99
I snuggled into the extremely comfy leather passenger seat of a Bentley as the sat-nav started talking, sounding a little bossy. The older man driving the expensive imported car is excited by the voice, a woman’s voice. He’s even given it a sufficiently British name. “I love it when Cynthia talks to me like that.”
My response was to laugh nervously. Now Tracey Spicer’s new book, Man-Made: How the Bias of the Past is Being Built into the Future, explains my feelings of extreme creepiness at the old bloke and the she, a disembodied dominatrix. Spicer puts it plainly: “The overwhelming majority of bots are built in the female form.”
Tracey Spicer presents a compelling picture of the way in which artificial intelligence reinforces stereotypes by replicating biases that entrench inequality.Credit: Janie Barrett
When men want slaves, they engineer them as women. Spicer’s new book explores the way technology reproduces all the bad times, just as tech leaders across the industry have called for a pause in the development of artificial intelligence. Last month Geoffrey Hinton, the artificial intelligence pioneer who created techniques that underpin ChatGPT, quit his decade-long job at Google so he could speak out about the dangers of AI. And this month, a statement from the Center for AI Safety, a nonprofit organisation, said: “Mitigating the risk of extinction from A.I. should be a global priority alongside other societal-scale risks, such as pandemics and nuclear war.”
There is no better time for this book: easy-to-read (suitable for school-age), racy, purposefully designed to appeal to mainstream readers and provide the link between the audience and the many academics researching AI. Spicer presents a compelling picture of the way in which artificial intelligence reinforces stereotypes by replicating biases that entrench inequality. Rubbish goes in and rubbish comes out, or, as Spicer explains it: “The misogynistic views of the past are being implanted in the news of the present.”
Be astonished – or not – at the multitudinous ways men take credit for women’s work in technology. As Spicer writes: “Women are involved in every step of imagining, building and maintaining computers.” They are the ghosts in the machine, argues Spicer, never recognised for their work.
When men aren’t taking credit, they build technology that works against women and affects the way women work. From the in-house thermostat set for the comfort of male bodies, to the phones built for the size of men’s hands not women’s hands, to algorithms that choose men’s CVs, to the voice assistants such as Siri that recognise and obey male voices more reliably than women’s voices, on earth and in space, the series of several striking anecdotes don’t quite make the case for a world built on male data in these early chapters – but you sense it coming.
Be terrified at the way Spicer reveals the extent to which we can be surveilled – not just by the state but by the people we think love us, the people we think we trust. We are victims of stalkerware.
Spicer’s approach is to cover a lot of ground in 300 pages, fair enough when you are writing a beginner’s guide and want everyone to understand the risks of untrammelled AI (rewards already well-touted) and how we got here. That means some aspects aren’t covered in sufficient detail, some claims are untested (I wish I knew of any academic discipline where researchers agree on boundaries).
I would have loved the opportunity to browse through more of the research – sometimes Spicer makes observations and it’s not entirely clear how she arrived where she did. The book should have solid end notes (it has some, but they are brief), many more references and an index – mind you, it might have been hard to convince Simon & Schuster to invest money in such an undertaking. I hope Spicer builds a website around the book and puts it all in there – mainstream readers want more.
Are we sufficiently alarmed about artificial intelligence and the way it is wielded in every aspect of our lives, including social media? The Algorithmic Justice League (AJL) says at the very least we should demand inclusive product testing and the regulation of auditing systems – but neither Meta nor Twitter is there yet. In the words of Ghanaian-American-Canadian computer scientist Dr Joy Buolamwini, who founded the AJL: “By the time we wake up, it’s almost too late.”
Man-made provides that wake-up call for everyone.
The Booklist is a weekly newsletter for book lovers from books editor Jason Steger. Get it delivered every Friday.
Most Viewed in Culture
From our partners
Source: Read Full Article