In 2016, award-winning journalist Tracey Spicer’s eleven-year-old son announced that he wanted a robot slave. So began Tracey’s six-year deep dive into the technological shifts that are reshaping our world. The resulting book, 2023’s Man-Made (Simon & Schuster), is a fascinating account of the historic inequalities embedded in the AI we use every day – and how we might overcome them. Griffith Review is delighted to publish a short extract from the book alongside a Q&A with the author. 

CARODY CULVER: You spent six years researching Man-Made, and I can only imagine how many technological developments occurred during that time that were relevant to the book. Did this change the trajectory of your research in any way?  

TRACEY SPICER: When I started talking about this idea in 2016, almost everyone said it wouldn’t work. They thought the book would be outdated by the time it was published. However, I always envisaged this as a work centred on the social sciences. It’s not really a technology book. Man-Made is about how humans interact with artificial intelligence, and the ways in which this constellation of technologies reflects and shapes society. Still, there was one significant development that we had to include during the editing process. When ChatGPT came out on 30 November 2022, generative AI exploded into the public consciousness. This was extremely fortuitous timing for the book! But it meant I had to quickly add some anecdotes related to the chatbot. 

CC: While Man-Made is a book about the future, it’s grounded in the past – you state how important it is to understand where we’ve come from in order to figure out where we might be headed. What’s the risk in thinking that artificial intelligence and the science around it are very recent phenomena?  

TS: This is one of the biggest dangers in the way we approach the topic. Humans have been making automatons since the time of the Ancient Greeks. Disruption due to technology is nothing new. We need to think critically about this fourth industrial revolution, remembering the lessons of the past. The rapid scientific discoveries, exponential technological advances and widescale job losses have all been seen before. To avoid worsening social dislocation, governments should consider a universal basic income. Interestingly, in the 1950s there were more women than men working in computing. The ‘IT guy’ phenomenon is relatively recent. We must remember this history to break down gender stereotyping in this sector and increase diversity and inclusion.  

CC: Man-Made is an incredibly readable book – the depth of your research is obvious, but you present it in a way that’s accessible and engaging (and you’re not afraid to crack a few jokes along the way). Was it important to you that this book appeal to the broadest possible audience? If so, why?  

TS: Thank you! Honestly, I loved nerding out on the research. It was fun going down various rabbit holes. But I’m not particularly technologically adept. And I didn’t want people like me to avoid the book, fearing it would be impenetrable. I worked very hard to make the science accessible, so anyone could understand the gravity of the problem. Working in the mainstream media for thirty-five years has taught me how to simplify complex concepts. There are some strong dystopian themes in the book, and shocking examples of bias and discrimination in datasets and algorithms. The humour softens the blow, somewhat. ‘Just a spoonful of sugar helps the medicine go down,’ to quote Julie Andrews in Mary Poppins.  

CC: Since Man-Made was released, have you come across any particularly promising developments in the field that may get us a step closer to the positive change you’re advocating for in the book?  

TS: There’s a promising movement known as ‘mindful AI’. This involves cleaning and diversifying datasets, auditing algorithms to remove bias, and using the principles of inclusive design. One of its proponents, Kieran Snyder, is the co-founder and CEO of Textio, a software company that creates inclusive recruiting tools. In New York City, employers must regularly audit automated hiring software for ‘bias against protected groups based on applicants’ race/ethnicity, sex and intersectionality’. The European Union’s AI Act provides a solid template for future regulation in various jurisdictions around the world. Importantly, people are using AI every day and beginning to see the flaws. Now, we need to turn that awareness into action.  

The following is excerpted from Man-Made: 

WE DON’T NEED a crystal ball. To predict the future, we must look to the past. The rise of new, urban manufacturing industries in the late nineteenth century fundamentally altered the nature of work. During the second industrial revolution, concerns were raised about child labour, working hours, pay and conditions. Most workers repetitively constructed components, deriving none of the creative pleasure that comes with producing a complete object. 

As artificial intelligence becomes ubiquitous, alarm bells are ringing about the rise of the gig economy fuelled by underpaid, overworked and invisible ‘ghost workers’. A twin threat is the huge potential for job losses in both blue- and white-collar work. Stanford University researcher Andrew Ng reckons the fearmongering about a dystopia caused by ‘evil AI’ – sentient robots intent on world domination – is diverting us from a much more likely outcome of this fourth industrial revolution: mass unemployment and its inevitable social and political impacts. 

‘AI is expected to be better equipped than humans to write a high school essay by 2026, drive a truck by 2027, work in retail by 2031, write a bestselling book by 2049, and perform surgery by 2053,’ Darrell West, vice-president of the Brookings Institution, explains in his 2019 book, The Future of Work: Robotics, AI, and Automation. If you want to write a book, you’d better get cracking! Before ChatGPT takes all of our jobs. 

Once again, people in low-status jobs – mainly young folks, people of colour and women – will be the most vulnerable. Governments will have to reskill workers and pay the unemployed to study. Education, the economy and welfare systems will have to be restructured. A universal basic income would be a robust solution: a government program providing every adult with an income to meet their basic needs. Historically, governments move too slowly to adjust to tectonic societal shifts. 

The first industrial revolution gave rise to capitalism. The second deepened the gap between rich and poor. The third enabled the transition to an interconnected economy. This works well when everyone’s playing together nicely. It falls apart when the US housing market sneezes and we all get colds during a global financial crisis, or Russia invades Ukraine and Europe suddenly runs out of gas. Mass unemployment would create a larger underclass. This is always a catalyst for social unrest, creating an environment in which extreme forms of populism flourish. 

Proto-fascism is creeping into politics, emboldened by white, middle-class people complaining they’re getting a raw deal. As we see in the US – and, at the risk of invoking Godwin’s law, in Germany under Hitler – certain groups do not fare well in these circumstances. If artificial intelligence turbocharges populism in politics, we face an apocalyptic future. 

People living in the global south will suffer the most. The AI economy is looking a lot like a colonial economy, in which people in poorer nations work long hours in low-paid menial jobs to provide profits for their Western masters. Women and children are already working in sweatshops to keep the West supplied with fast fashion. Will artificial intelligence entrench this kind of economic colonialism? 

At issue is the lack of diversity, the disparity in power and profits, and the structures continuing to exclude women and people in marginalised communities. An industry that loves to see itself as ‘disruptive’ is reluctant to disrupt the male-centric, capitalist and colonialist structures that threaten technology’s potential to create a kinder, smarter and healthier world. Ironic much? 

In future, we will need a major redistribution of wealth. Vested interests would fight tooth and nail to maintain the status quo. Mega-companies like Amazon, Facebook and Google – which are involved in a digital arms race – would have to be broken up. Tech billionaires would have to decentralise systems that currently funnel AI profits to individuals.  

In 2017, Bill Gates tells Quartz that governments should tax companies’ use of artificial intelligence robots to temporarily slow the speed of automation. Gates rejects self-regulation, insisting governments must oversee programs of wealth redistribution. We all know the dark side to the ‘boy genius’, but this idea is brilliant. Believe it or not, it’s even considered by the European Union. Sadly, European legislators ultimately reject the proposition. 

Let’s take a deep breath: dystopia is far from inevitable. But almost without exception, the experts I consult are very worried about the road we’re travelling. So how can we move towards a more utopian vision? First, we must ensure our AI future is humanistic, collaborative and intersectional. Women, people of colour, the LGBTQI+ community, those with disabilities and older folk need quite a few seats at an awful lot of tables. Seismic changes are required at schools, within industries and at the local community level to support more inclusion in STEMM and STEAM. 

People from disciplines beyond STEM will have to be involved in planning and developing artificial intelligence systems, especially in terms of identifying and managing risks. ‘AI is a bit like maths in a way, it’s going to be required in every discipline, and so we need to be training up future workers not just to be experts in health or law or manufacturing, or whatever it is, but to have those underlying digital AI skills as well,’ Data61’s Jon Whittle tells the ABC in 2022. ‘There’s a wonderful quote…in the past jobs were about muscles, currently they’re about brains and in the future about the heart.’ 

But what does this mean for specific sectors? The ability of artificial intelligence to mitigate climate change should be explored, but the industry must take steps to reduce its own footprint. This technology can be a boon for women’s health and safety, but devices need to be designed in consultation with diverse groups. Health care of the future will be predictive and preventative, rather than reactive. But it has to be accessible to everyone – not just white people in wealthy countries. AI could change the lives of people who use social services. But the users must be included from the outset. Innovation has the potential to smooth every aspect of our lives, from shopping to transport, employment and household duties. Machines that gestate human life could free women from reproductive obligations and biological restraints. A step too far for many, but something to think about, nevertheless. If developed sensibly, artificial intelligence could take part of the burden of caring for children, people with disabilities and elderly relatives away from women. 

Currently, Indigenous people are looking into how artificial intelligence can help conserve traditional lands, revitalise lost languages and preserve ancient knowledge. This is through something called artificial-intelligence-enabled cyber-physical systems (AI-CPS). Put simply, these systems could monitor wildlife, reduce poaching and simulate ecological interventions. Nowadays, Indigenous peoples manage around a quarter of the world’s land surface. Imagine the impact of combining old and new knowledge. It could sustain the land and its inhabitants for future generations.  

In Australia, artificial intelligence is already being used to care for Country. During his Master of Applied Cybernetics at the Australian National University, Rodolfo Ocampo works on the CSIRO Healthy Country AI project. Developed in collaboration with Bininj Traditional Owners and Indigenous rangers, the project manages invasive weeds, which are decimating biodiversity in Kakadu. 

‘The Healthy Country AI-CPS uses drones to survey the land, machine-learning algorithms to classify aerial images and cloud technology to visualise infestation maps,’ Ocampo writes on the Australian National University website in 2021. ‘Algorithms and data management were trained integrating traditional knowledge. For example, the Bininj six-season calendar is used to classify drone footage.’ We have so much to learn from Indigenous culture, including in the area of artificial intelligence.  

We need to work together to create a utopia. This begins with words and ends with actions. ‘I tend to think we have an obligation to tell stories about a future that is more just and fair and equitable and sustainable, and thus also more optimistic,’ according to the chancellor at the Australian National University, Distinguished Professor Genevieve Bell. ‘And I think we also have an obligation to actively disrupt the present to make those stories possible.’ 

This time in history is nothing short of transformative. However, it’s painful and exhausting because we’re evolving from one stage to the next. ‘Are we in the process of giving birth to a whole new world?’ social commentator Jane Caro writes in The Sydney Morning Herald. ‘Hopefully a fairer, less hierarchical world, where we will live in greater harmony with one another and with the natural environment? Is that why we are in such an uncomfortable and panicky place right now? Just like mother and child, we know the world we are leaving but we do not know the one we are going to.’ 

Maybe we should all go with the flow, confident of soon holding new life in our hands. But it takes more than one party to create life. That’s why the future must be made by all of us, for all of us: human-made, not man-made.