Skip to main content

Capitalism Drives the Turing Trap


In The Turing Trap: The Promise & Peril of Human-Like Artificial Intelligence Brynjolfsson looks the risk AI poses to jobs, and hence to society, concluding that using AI to enhance human labour is a better course of action then eliminating humans from the workplace.

He notes a drive to develop human like intelligence although most forms of AI are currently not human like. Developing human like AI will concentrate economic and thus political power in whoever controls the technology. He does not consider the possibility that human like AI might become conscious or superintelligent and end up controlling everything.

He does however seem to acknowledge that business is about power relations as much as profit. A business owner, in my opinion, is likely to be primarily concerned with profit (and economic or political power) while managers, at lower levels, whether in Charles Handy’s verticals or horizontals, will be increasingly concerned with power over their subordinates and the lower they are in the hierarchy the more they seek power.

He notes that either replacing or augmenting humans can be beneficial, but augmentation allows human creativity and creates now products and services. There are however a number of perverse incentives for technologists, business leaders and policy makers (politicians) to favour replacement over augmentation. Technologists love the challenge of replacing humans while business owners see a reduction in costs and therefore use their economic power to steer politicians to favour replacement.

The Turing Trap is the drive towards replacing humans at the expense of creativity and innovation and the resulting concentration of power in a techno-business complex.

Technologists are trying to do more than just replace humans, and seem to be slowly moving towards replacing nature. for example,by Developing Robot Bees and have designed a robotic chameleon that crawls and changes colour.


Automation Versus Augmentation

One major benefit of automation is elimination of mistakes. Early mathematical tables, for instance, were plagued by errors which were eliminated once calculating machines such as those invented by Charles Babbage, could automate the whole process from computing the figures to making printing plates.

Another benefit if automation is doing dirty, dangerous or boring jobs, especially where boredom is likely to generate mistakes by human operators.

It is unclear whether replacing our current politicians with Human like AI would result in better governance: that would depend on the quality of the specification and on the quality of the reviews of the finished product.

Scroll to Continue

The automation/augmentation dilemma is therefore not an either/or dilemma but a question of which is best in particular situations. Driverless cars are almost certain to make travel safer, but a robot secretary or accountant is likely to be less effective than a human with an AI based personal assistant.

When technology is used to augment human abilities it creates a new case of skilled and hence expensive workers. In past centuries these might have been printers or Morse Code operators. With time automation deskilled these tasks and allowed businesses to replace these workers with machines operated by cheaper commodity labour. We can expect that initially augmentation will create new castes of skilled workers but they will later be replaced by domain specific AI guided by or guiding relatively unskilled labour and eventually the human components will be eliminated.



Business people see a reduction in wage costs from automation but externalise the sociatal damage that would be caused by mass unemployment if replacement became universal. They also forget that customers have to be able to afford your products. If robots replaced all jobs so there was no need for human labour the machine owners would not be able to sell their products and services. Clearly there is a point at which the unemployment caused by AI would cause economic collapse, perhaps even social collapse.

It is possible however to imagine a totally automated economy and society in which every need was provided for free by intelligent machines. This would not eliminate work, merely paid work. Art and entertainment would persist. It would merely be businesses, the owners of the technology, who might vanish or change in ways we cannot foresee.

Human Like Intelligence

Human like Intelligence would be able to “think on its feet” responding to and learning from unforeseen circumstances perhaps being able to determine that an apparent fleet of incoming missiles on a radar screen actually represented a system malfunction thus preventing a war. Businesses will face pressure to replace humans with human like AI once the AI exceeds human abilities on specified tasks and the political and fiscal environment currently favours such replacement and pushes the evolution of technology towards replacement of humans.

Avoiding Replacing Humans

Augmentation and Automation can both be beneficial but automation concentrates power in a smaller number of hands and there is no certainty that those without power will gain any benefit or even keep what they have.

This is particularly true in the case of an unfettered market, something more closely approximated in the slavery era than today, and in the times when scrip payment was legal.

Law and lawmakers, as can be seen from the case of slavery and scrip payment, have a role to play in alleviating the pressure to replace humans with AI rather than augmenting human abilities.

There are strong fiscal pressures to replace humans with machines. Payroll taxes such as employers national insurance, holiday pay and workplace pensions vanish with machines replace people and a preferential tax regime for investment in capital rather than people. The situation could be reversed by an increased corporation tax and adjustment of incentives. It could even be tax neutral. If this is to be done. It should be done now before the shift to replacing humans becomes irreversible.

In Conclusion

The trend to replace humans with machines is driven by perverse incentives and can be reversed by changing these incentives. Augmenting humans gives power to workers and encourages engagement with their work. Automation replaces humans, concentrates power with the technology owners and prevents the workers benefiting from the replacement.

There are however some situations where automation is necessary: dirty and dangerous jobs or jobs where human error may be disastrous.

In other situations the relative merits of these two strategies for the business, the workers and for society need to be considered. The perverse incentives pushing for automation must be removed so an unbiased choice can be made. This is unlikely to happen as long as business can buy politicians.

Related Articles