AI won’t take your job, prosumers will
There has been a lot of talk about AI taking away jobs. Maybe the most high profile is Martin Ford’s (@MFordFuture) Rise of the Robots. The argument is that more sophisticated Artificial Intelligence will make it possible to automate jobs that are currently done by people, and so current jobs will disappear from Truck Drivers to Radiologists and more.
The advances in Deep Learning do certainly seem to point to a much more powerful AI, capable of many more things that even recently we would have considered jobs for humans, not computers. But if we look at the recently history of automation, the picture looks a bit different.
Over the last 20 years technology has already replaced or at least reduced a lot of jobs, from bank tellers to travel agents. These jobs can be seen as being the results of automation, but they are also something slightly different. Bank tellers have been replaced by Automatic Teller Machines (ATM, or cash machines as we call them in the UK) and online banking. What these technologies do is enable customers to do for themselves what was previously done by people. The same goes for travel booking. There are dozens of web sites that let you find and book flights and hotels. The jobs aren’t being simply replaced by technology, they are being done by consumers. This is even more obvious if you think about YouTube replacing (some of) our TV watching with user generated content (not to mention medium).
Prosumers
In his amazingly prescient 1980 book, The Third Wave, Alvin Toffler predicted a future in which life would no longer be strictly divided into work and leisure or production and consumption. Instead everyone would take part in production while consuming, they would become what he calls “prosumers”. He says that this harks back to ancient and medieval times where people would have made their own clothes and other possessions. It was industrialisation that turned us into producers and consumers where producing things was a specialist form of work done by a few, while everyone could consume the results. Toffler predicted that the era of this stark divide was ending, and that technology would enable everyone to take part of the production of the objects they consume. This seems pretty reasonable now when we think about ATMs and YouTube, though I still find it stunning that he was writing about this almost 40 years ago.
So our future does seem to be one of prosumers, where technology enables us to do work and creating things ourselves rather than depending on professionals.
Prosuming AI
What does this mean for AI? Will AI change this direction of travel away from prosumption and towards complete automation? Possibly, AI could change the trend of the last 20 years. But I think that it will accelerate the trend. While some things like autonomous trucks could be seen as complete automation, I think that the combination of humans and machine will remain more powerful, humans can do things that AI can’t.
One of the things that has become obvious with the development of AI, is that often the hard things for AI are not the things that are hard for humans, or need us to be advanced experts (grandmaster chess is the classic example). It is just as often things that almost anyone can do (walking down the street without bumping in to things). So rather than complete automation, AI could remove the need for some expertise and enable anyone to perform complex tasks, exactly what we need for prosumers.
AI is already enabling prosumers. There are AI services with which you can proofread your writing. The quantified self movement generates a lot of data that is ideal for machine learning. Off the back of it there is AI that can help you be your own personal trainer and there are starting to be AI systems that help you self manage health conditions, something that is likely to be a big part of the future of medicine.
I’m a university educator. I think it will be a long time until AI is going to replace teachers, but MOOCs are enabling people to prosume education (something I personally am enthusiastically taking part in). They even include learners marking each other’s work, something that is a great learning experience. Social MOOC learning means that people can learn from the experiences of other learners. MOOCs and online learning are increasingly using AI to support learners. MOOCs generare masses of data about learners, which can be used with machine learning to provide personalised support. I personally think that this kind of data driven learning will work best if learners can have access to their own data and the output of AI so they can make informed decisions about their learning, rather than AI making decisions for them.
Human-Centered Machine Learning
If AI is going to be able to support prosumers it isn’t enough to be smart. It also has to be usable. In all the examples I have listed the interaction design is at least as important as the AI, if not more. We need to move away from the idea of an autonomous AI with no human input, and think about how humans interact with AI: the HCI of machine learning, or something my colleagues and I have called Human-Centered Machine Learning. Google is a great example of this, they have some of the most impressive machine learning on the planet, but it is all in the service of a simple and elegant user experience. I believe that the combination of interaction design and AI, not AI alone, is going to be what drives the future of technology and even of our whole society.
The future of work
So what does this mean for the future of our work and lives? The optimistic viewpoint is that we will all have more control over the products we use and we will all be able to be more creative and have more effect on the world as we move away from the hegemony of big content producers. The pessimistic view is that we will all loose our jobs and what is worse we will end up working for free.
There is always some of the good and the bad in the final result. I believe that more prosuming in education will open up education, and lifelong learning, to people who would not have had access to education otherwise. That will be about expanding the whole education sector, not about replacing jobs. I think the same will be true of medicine. The same number of doctors or professors will be able to reach far more patients or learners.
But there is likely to be less paid work over all. That will be accompanied by more upaid leisure activities, our prosuming. To me the answer is obvious, though very few people are talking about it. We need to work less. The smaller amount of work will be spread among more people. It will leave time for prosuming, whether it is creating youtube videos, or managing your health or that of your family. We might all even have a nicer, less stressed life. Working less is a solution that seems almost inconceivable in our current society, which is based on everyone working harder and harder, and I have to admit it is hard to see how we can get there. But I think it is something worth aiming for. One way of thinking about it is that as more things are automated and prosumed, more will become much cheaper or free. That means we could work less and earn less without losing our quality of life, possible improving it. So I think there is scope for optimism about a prosumer society, but we need to start talking about working less, which is completely absent is current politics, but is a completely logical consequence of the development of technology.
