Defend Truth

Opinionista

Machines take over: It’s time for a complete overhaul of Marx’s theories of modes of production

mm

Professor Tshilidzi Marwala is the seventh Rector of the United Nations (UN) University and UN Under Secretary-General.

While I always thought that there were many possible patents that people could invent, particularly with the advent of artificial intelligence, I never considered that there could be patents invented by AI itself. Now South Africa has become the first country to register a patent that was created by an AI machine.

More than 10 years ago, my collaborators and I registered a patent in the United States for our invention, which uses artificial intelligence (AI) to create an artificial larynx. The larynx is a vital human organ responsible for our voice. Typically people lose their voices when the larynx is surgically removed because of cancer.

Our invention was called the Robot Voice by the Massachusetts Institute of Technology (MIT) Technology Review. While most devices leave the speaker with robotic or monotonic voices, our device reconstructed the intended speech through a speech synthesiser to achieve a voice pattern as close to the speaker’s original voice as possible. Naturally, with an invention such as this, we wanted to protect our intellectual property. We were able to do this by registering for a patent, which is a commercial privilege that is given to inventors to protect their inventions from infringement or use without the inventors’ consent.

While I always thought that there were many possible patents that people could invent, particularly with the advent of AI, I never considered that there could be patents invented by AI itself.

Last week, South Africa became the first country to register a patent that was created by an AI machine. The invention was an interlocking food and beverage container formulated on fractal geometry. Commenting on the patent, Dr McLean Sibanda, an expert on intellectual property, said “this patent has been refused elsewhere in the world where patent applications are subjected to substantive examination before being granted. Main grounds have been that AI is not a natural person and thus cannot be an inventor, and as such cannot duly assign its rights to an invention to anyone to apply for a patent.”

As this case has demonstrated, AI is increasingly assuming a human dimension. It is only logical then that the laws that prescribe that only human agents can acquire patents are reviewed and updated. Australia is another country that recognises a non-human inventor.  

This is not the first instance of AI replicating the kind of innovation that has long been considered only a human capability. In 2018, 1 the Road was the first novel written by an AI machine. Given that it was not a human being who wrote this novel but an AI machine, is it not logical for the authorship to be attributed to the machine? If we can readily assign authorship to an AI machine, why can’t we assign a patent to an AI machine?

This era of intelligent machines performing tasks that contribute to production independent of humans heralds what Calum Chace calls “economic singularity”. Chace defines this as the point when AI “renders most of us unemployed, and indeed unemployable because our jobs have been automated”.

Singularities were first predicated as a result of Albert Einstein’s Theory of General Relativity, which entailed a point in space-time where the physics laws as we understand them break down, implying the existence of black holes. The central characteristic of a singularity is unpredictability. Economic singularity implies a mode of production where machines surpass human intelligence to such an extent that they become the primary drivers of economies. They are thus able to make other machines and produce goods and services with hyper-efficiency. 

The mode of production is a concept conceived by Karl Marx and is composed of the means and relations of production. The means of production are the tools or machines that are used to produce goods and services. The relations of production are the relationships that exist between the people involved and the instruments used in production.

For example, in the slave mode of production, the slave master owned the means of production, including the slaves and worked these slaves to death to extract surplus value or profit. 

In feudal society, the mode of production entailed landowners extracting profit from the serfs who worked for shelter and food. In the capitalist mode of production, the working class creates surplus labour (profit), which is exploited by the owners of capital (means of production). Marx called the workers the proletariat, and the owners of capital the bourgeoisie. To deal with exploitation in the workplace, Marx proposed abolishing private property in favour of state ownership of the means of production. The Soviet Bloc attempted to abolish private property, but failed spectacularly.   

In the era of economic singularity, a new mode of production is emerging right in front of our eyes. In this mode of production, the means of production is increasingly marginalising labour. Consequently, machines are replacing labour in production, leaving owners of the means of production and AI machines as the only forces of production.

Various thinkers have used different terminology to describe the era of economic singularity. Some people have used the terms hypercapitalism and the Fourth Industrial Revolution.

Whatever the terminology we may use, in this era the resulting relations of production will increasingly be between owners of capital and automated machines used for production. These machines will be so sophisticated that they will invent other machines. The AI machine that registered a patent illustrates this issue. They will be able to predict the state of their health and fix themselves.

What is to be done in this new era where machines are exceeding our capabilities? Abolishing private property in favour of nationalising the means of production to protect people from being impoverished is not the answer.

There are two things that should be done.

First, these intelligent machines should be taxed. Second, the proceeds of these taxes should be used for a universal income grant. Anything short of this will plunge our world into depths of poverty and inequality from which we will not emerge. Strategically using these intelligent machines that are capable of owning patents can reverse this fate and could end poverty in society. DM

Gallery

Comments - Please in order to comment.

  • Steve Spottiswoode says:

    Be very careful with any Universal Income Grant not to let it grow so high that people choose not to actively contribute to the country or area they live in. The same could of course also be said of any form of income!!

  • Retief Joubert says:

    This is the umpteenth article by the good Prof where I have to wonder what world he lives in. We cannot keep the lights on, but ja, lets talk about the taxation of AI. Same Prof who sprouts the wonders of the 4IR looming on the horizon, yet our education system spews out illiterate, unemployable matriculants by the tens of thousands.

    Your vision(s) will not see the faintest glimmer of day until govt gets the non-negotiable basics right for a decade or two. Stop stealing. Start working. Appoint on merit.

Please peer review 3 community comments before your comment can be posted