Towards a Project-Based Economy: AI and the Future of Work


An AI lawyer is doing legal research in the US, a robot is laying bricks in Japan and a robot just passed a visual Turing test at MIT. The question is no longer whether automation will occur but how long we have to control its introduction and the future of human work before it does.

While the media continues to argue over what the utopian future will look like when robots do our laundry and dishes, a much bleaker picture is emerging of humans working unregulated working hours in labour camps under crippling working conditions to ‘compete’ with their robot colleagues in the workplace. Robots are estimated to potentially work for just a penny a day, making them much tougher competition than the supposed threat of low wage immigrants.

In this article, I intend to argue against the popular misconception that robots and AI will lead us towards an unprecedented revolution in individual freedoms.


Since the Second World War the story has gone in the opposite direction, as technology and automation has been used to make workers work longer hours. The anthropologist David Graeber points this out. Despite significant changes in technology in the last few decades there has been almost no reduction in working hours. Instead, our economy has gone in the opposite direction, creating ‘filler jobs’ or ‘bullshit jobs’ as he calls them, to fill in the vacuum of dying automated industries. Where automation was meant to free up people to do what they like with their time, it has instead been used to make us work longer.

The average full time working hours in Australia remain at forty hours a week, with around seventy percent of the economy working a full time job. The majority of the population do not work in food, clothing or shelter but in ancillary “luxury” industries like advertising. Jobs in advertising, marketing, social media and so on have seemingly been invented for the sake of keeping us working longer hours, as opposed to any direct survival-based output. They hardly have an evolutionary imperative, but rather an imperative of social continuance – of people being retrained and reskilled to keep them working.

This process of inventing jobs to keep people working is at the heart of why AI and automation may not free up our lives in the future, even with the resurgence of robotics. Whenever automation has been close to granting people individual freedom – a new industry has risen up to keep people working the same hours, indefinitely.

Our politicians are already talking about AI in the same way as they discussed previous introductions of automation. They talk primarily of training young workers for “jobs that don’t exist yet,” for an “uncertain future” and so on.

When they admit that automation will eliminate certain white collar industries, they call for a reskilling of workers to fit into “new industries,” yet to be created. This is coded language for the idea that working hours will never be decreased regardless of how much the economy gets automated.

Instead of reducing working hours new industries will be created to force humans into servicing robots and AI machinery. The recent focus on “STEM skills” and “coding” in this context is particularly pernicious. There is a reason why ‘coding’ is such a buzz word, and it has nothing to do with politicians liking to code. Facing the prospects of a 40% unemployment rate, the government can do little but force young people into narrower and narrower specializations with regards to robotics and technology, in an attempt to prepare them to create “future industries” using “STEM skills”. Indeed, a recent Oxford study predicted that around 35% of current jobs are at risk of automation over the next ten to fifteen years.[1] The remainder will be technocratic positions to manage the fallout from automative work.

If this all seems farfetched, a recent example should give us reason to pause. Perhaps the most famous example of automative technology in the last century, the computer. Both the computer and the internet were once heralded as the main harbringer of individual freedom. They were expected to lower working hours. Computers were expected to free up our time while drastically increasing our productivity.

In reality, productivity increased but working hours remained static – wages flatlined and living conditions for the working poor declined dramatically. Blue collar workers in particular suffered declining standards of living, marking decades of decline.

Despite computational technology, or perhaps because of it, we are working longer hours than ever before. Employers can now contact their employees outside traditional work hours– by email, text, social media and so on. Instead of freeing up our time to spend on projects of passion, automation over the last two decades has been used to make us work 24/7. Erasing the very concept of a ‘private life’. The problem is so bad that the French government recently banned employers from emailing their staff after office hours.

The story wasn’t meant to go this way.

In 1933 the economist John Maynard Keynes famously predicted that in the 2030s humans would have a 15-hour workweek due to massive advances in technological innovation and change. Keynes prophesised that people in the 2030s would ask not what to do with one’s work but what to do with one’s life in general. Human purpose would replace money as the central question of value in society leading to a new Renaissance of individual liberty and purpose.

Instead of preparing young people for “jobs that don’t exist yet,” political leaders should learn the lessons of history from the creation of the computer. Instead of working the same or longer working hours, AI should be utilised to replace human labour entirely – allowing for Keynes’ 15 hour working week to be realized.

While many are suggesting a basic wage as a possible silver bullet for the problem, it is clear that at a minimum a complete re-think of the nature of work is in order.


For another interesting take on AI replacing human employment, see: