You like your job. Yes everyone grumbles, but doing a good day’s work and being paid for it feels good – and it pays the bills. But today, if you peek past your computer into the world beyond you’ll see the vast tide of automation approaching – devouring careers and and leaving nothing in it’s wake but unemployed masses and huge profits for business owners.
I want to be positive in this article, because that is the nice thing to do. Yet Automation is inevitable, in the same way as mechanisation was. It has obvious benefits. Goods will be cheaper to produce, and of higher quality. There will be fewer workplace accidents, and that will save lives. The problem with it, as with all prior human revolutions is that it will create huge inequalities which – unless handled deftly – are likely to last beyond our lifetimes.
Lets say you’re one of the lucky few who doesn’t have to immediately worry – your work involves non-routine tasks which require social intelligence or critical thinking. The few things we are pretty sure software isn’t great at – yet. Chances are you can retrain and supervise or do something else that you love. For the rest of us that option isn’t open.
We want to be drive trucks? Nope. Driverless cars take that off the table along with almost everything else. Medicine is increasingly supported by big data systems, data science and AI. Architecture is largely governed by regulations and so can be automated aside from the occasional creative flourish within the bounds of (digital) reason.
Even the lawyers should watch out as contracts are now routinely read, and analysed, and cases prepared by AI’s. Politics itself is a matter now of having the right advice, and the right decisions are almost always backed by at least some kind of software modelling. We are beginning to rely on the machines we build to decide as well as do, and that won’t change as the machines improve.
The estimate is that 44% of ‘low skilled’ jobs will be eliminated by the mid 2030s. So in 15 years most of the workforce will either have lost their jobs or will counting down the day to when their chosen career will be permanently over.
In the UK PwC studied the trends and determined that Automation will actually happen in waves. They identified 3:
- Algorithmic (upto early 2020s)
- Augmentation (upto late 2020s)
- Autonomy (upto mid 2030s)
In plain English the first wave – Algorithmic – which is in full swing right now – automates structured and standardised digital tasks. You can see it in automated phone responses, financial services algorithmic trading.
The next wave is when AI automated services Augment decision makers through most sectors where the quality of decision making is governed by the relationship between information and results (basically all sectors of the economy, think AI business owners, lawyers, doctors, and other professionals). They may still be the doctor, but it will become increasingly less likely we’ll go with the fickle human when the reliable robot can deliver better survival rates, greater profits, and more wins because quite simply they don’t miss things that us frail humans do.
Finally comes true Automation, when AI has matured to the point of anticipating future needs in the economy, society, and politics, and provision the required resources and strategies to make them a reality. A lot of it will be like a dance of dragons, the AI’s will either make space for each other or they will battle it out and inevitably some will win and some will lose with human beings picking up the final costs.
What happens beyond this anyone’s guess. So where does that leave us here in 2019? Can we retrain, what are the likely effects?
The fact is unless there is something place like Basic Universal Income the inability to compete with machines for leave a lot more people poorer. The more educated and flexible, the younger and more adaptable of us will try to create new niches to survive, however you are going to see the continuing concentration of wealth into startlingly few hands in the course of our lifetimes.
That will leave the rest of us as consumers who can’t afford to consume, and workers who can’t compete to work. Under current social and political conditions that inevitably means people will suffer, starve, and yes a lot of people may die. There are likely to be riots. Counter movements. Concessions yes, but history is not going in the direction where the rich give up the chance to be richer. With machines to enforce their will they can see a day where labour problems, immigration and social diversity simply cease to be issues they care about.
Put simply Automation will be the revolution of the elite, for the elite, by the elite.
There are ways around this. The first approach is to decentralise the economy, by moving away from traditional fiat systems to those which aren’t reliant on any single individual or political entity. This breaks the stranglehold of existing hegemonic and political entities – and we know that isn’t an easy path given the existing disparity in power. If that way is to succeed a decentralised platform independent of monetary influence will be a necessary first step, then perhaps a decentralised internet followed by decentralised computing services free for use at the point of delivery. This technology goes hand in hand with encryption and is unlikely to garner great support from the status quo.
Follow the golden breadcrumbs!
The second path is to use existing computing services to model societies and try to design economic systems and find a ‘Golden Path’ through automation which doesn’t worsen the problem of inequality. A combination of Basic Income, with free training, a decent social welfare network, and the anticipation and preparation for new economic niches could prevent much of the downsides and allow the population to migrate to the new economy where our robot workers are seen as allies rather than competition.
There is a strong argument that increased taxation is the wrong approach and that wealth should be re-contextualised in an age where even the executive talent of the wealthy is due entirely to the success of machines smarter than them. A person may own a megafactory of AI workers, but those workers will not make any money if no one can afford those products, and they wouldn’t have a place in most people’s lives. Market forces cut both ways and we have learned that the hard way in the past few decades.
Legislate the danger away.
The final way is to put legal restrictions on AI to prevent Automation from entirely transforming the economy. AI’s can be as smart as they want, but they can’t be networked, or use the Internet. Perhaps there should be limits on how many AI’s a person can use, or how smart they can be. Almost the entire discussion around AI for the past few decades has been on their potential danger to humanity and we should perhaps take that seriously and restrict their ability to overtake us. Adding AI’s to the Geneva convention is a good start as well. Perhaps business should require a licence before they are allowed to automate which ensures that jobs are migrated safely so that individual populations don’t get impoverished. This approach is less analytical than the ‘Golden Path’ however it is much more feasible and likely to avoid a great amount of the anticipated fallout.
Whatever happens we should welcome the benefits of smarter AI, and increased automation, while being mindful of the costs.