On 10 December the IWI Symposium on “Artificial Intelligence and the Trade Union Movement” was successfully held through ZOOM. More than 90 trade unionists from 35 countries participated in the seminar.

The keynote speech was given by cde. Grigoris Lionis, Dr. Mech. Engineering. An intervention was also made by cde. Sudip Dutta from CITU India and an expert on nanotechnology, on the topic “Artificial Intelligence: Utopia or Dystopia?”.

See below for presentations, speeches and photos from the Symposium:

 

Artificial Intelligence and Trade Union Movement

Grigoris Lionis (Dr. Mech. Engineering)

The debate about the notorious artificial intelligence, AI, and its effects on humans, and in particular its effects on workers, is a debate that has recently flared up.

Almost every day, dozens of news and analyses about new developments in AI, new applications in production and how it can make people’s lives much better are brought to light.

And of course, the workers’ movement has to discuss this issue, to shed light on its aspects, to answer the questions raised, to anticipate the future developments.

Our discussion today aims to make a small contribution to this process, to trigger new thoughts and viewpoints.

Before we look deeper into the issue, and try to give some answers as to what AI really is and what its limits are, we have to start from the understanding that AI and its application in production is a reality and concerns practically all workers on the five continents.

The examples of how AI impacts the lives of workers around the world are numerous:

• In India, there are various announcements about AI call centers , their success rates and their ability to automatically handle the vast majority of calls.
• Brazil is promoting a major plan for the development of AI, which in addition to its use for “public services” is also related to industrial policy, with Minister Márcio Elias Rosa stating that the government’s goal is to “put AI inside the factories to make Brazil efficient and competitive”.
• In the US, AI applications are being used in an increasing number of companies in a growing number of industries, while the motto “bring production back to the US” is being utilized.
• The EU is moving forward with many new laws to regulate AI while “racing” to catch up with competition in its implementation in the production process.
• China is a world champion in the implementation of industrial robots.

And so on and so forth.

Hence, a number of questions are raised by the developments themselves. WHAT is AI and, above all, what can it do? How will work with AI evolve? Should we be technophobic or technophiles? What should the workers’ movement do in the age of AI? How can it be utilized for the benefit of the peoples, for the benefit of workers?

Much can be said about AI technology, but it is worth focusing on some facts that can promote the objectives of our discussion.
The term AI is more or less an umbrella term that encompasses many, different technologies and technical solutions, which nevertheless converge on one result: the AI function seems to replace human cognitive functions. Certain processes that until recently required a human mind to perform can now be performed by computer AI systems. The processes that are being automated relate both to activities that are mainly in the mental realm – e.g. translating a text from one language to another – and to activities in which mental functions control physical functions, such as driving a vehicle, or building a wall, which require not only physical strength, but also the ability to perform the tasks of the human brain: perception of the environment, understanding of the problems that the environment can “create”, choosing a different approach etc. The older automation processes – automated industry, traditional robots – worked well in a standard environment, strictly defined, on a production line, and even there they had little room for flexibility, making their application feasible only in very large volume applications. AI “gives” these systems a much greater flexibility. After decades of development of such systems, mankind has managed to build systems that can now operate in such conditions, and replace humans in some of these operations.

These systems use huge computers, purpose-built for the job, and unlike traditional solutions they need to be “trained” with a very large amount of data, which are primarily human-generated.

In this sense, these systems actually have a social basis since their “intelligence” is, after all, nothing more than the “distillation” of the intelligence of the people with whose data these systems are trained. Simply put, the great language models that “speak”, in fact do nothing more than “parrot” relevant words and phrases, and compose sentences, paragraphs and texts that look like something they have already worked out. They have simply edited tens of millions of pages of text, which they practically remember all at once.

Thus these systems, always in their current version, do not “think”. They do not understand what they are saying. But they produce such complex responses that it seems as if they were thinking, as if they understood what is being said.

This is why their ability to replace humans altogether is currently non-existent.

However, these systems can – in the way they work – multiply the productivity of labour, in such a way that a significant reduction in the number of workers is allowed, as workers with AI systems perform tasks that were previously performed by many more workers. The recent example of a large call centre company revealing that automated call routing is used in 95% of calls is telling.
In the coming years, we can expect rapid development of many such technological systems that can perform processes with a significant reduction in manpower, from robotic drivers performing 90% – 95% of the journey, to robots in the industrial production process.

Will human labour be eliminated with these systems? Speaking of the medium-term future, the answer we give is a resounding NO. These systems lack awareness, judgment and the ability to understand and respond to truly unexpected developments.
However, the fact that the productivity of labour has soared with the use of such systems, which allow tasks to be carried out with a significantly smaller number of workers, is clear and beyond doubt.

Almost all studies assessing the impact of robots and AI on the workforce conclude that the adoption of robots leads to a loss of jobs in the industries that deploy them. In fact, these studies refer to “simple” robots that are already used in industry. The new robots operating with AI systems, which are coming in the near future with even greater capabilities, will have an objectively greater impact.

This phenomenon, of course, should not be surprising to anyone, and after all, it is not new. In every phase of development of new technological solutions, the new level of automation has been used to increase the productivity of labour and to replace human labour in one sector that was then transferred to other sectors.

Perhaps the most typical example is the example of agricultural production. Before the mechanization of agricultural production, the latter absorbed a very large part of the labour force almost all over the world. Mechanization, the automation of agricultural production, drastically reduced the need for labour force employed in it, and historically freed it up to go to other sectors of production.

The technical difference with AI possibly concerns the fact that the changes, the automation of production and the tendency to replace human labour concerns almost all sectors of the economy at the same time. The advent of AI does not concern one sector or another, but instead concerns all work sectors, or at least a wide range of them.

But above all, what influences the effects of the new technological means on work and human relations in general is the social structure itself, capital and the ownership of capital.

When studying the effects of new technological means on work, we should bear in mind that in capitalism, the means of production constitute capital. Capital is invested with the sole criterion of its self-expansion, its profit in simpler words. Thus, capital investment, the whole of investment in capitalism, is promoted on the basis of profit and the rate of profit. Investment is prioritized on the basis of the expected rate of profit. In other words, new means of production are developed and used according to the rate of profit.

In the area we are currently examining, in new technologies, AI and robotics, this observation explains how capitalism will use the new means of production and the increase in the productivity of labour. Investment in the new means of production occurs when the “cost” for the investor is reduced, i.e. when either the labour force required in the particular firm is reduced – and indeed when it is reduced enough to compensate for the increased investment – or when the new means of production expands market share, thereby reducing employment in other firms in the industry.

A relatively recent study carried out by the Massachusetts Institute of Technology (MIT) in 2020 found that each industrial robot installed in the US reduces – in the wider region – the workforce by 6 workers. The effect in other countries with a different economic structure may be different, but the fact is that investments of capital, i.e. the investments that capitalist states and companies make in robots and AI have the following effects: they replace human labour, reducing the number of jobs, while changing the jobs that remain, making work more complex, often longer, and, due to unemployment, they can even result in wage cuts for those who work.

Moreover, as we examine developments within the context of the capitalist mode of production and today within the context of monopoly capitalism, new investments in production automation, robots and AI have the potential to affect even geo-economic aspects of production. Some of the cheaper labour employed in developing regions of the world may be replaced by automated means of production in the richer, consumer regions. To give a simple example, an automated system that produces clothes, drastically reducing the need for direct labour, could lead to a significant removal of manpower from SE Asia, which is currently the “matrix” of garment and footwear industry.

Apart from the economy, these new technologies are also critical for political and military power. We see new technologies being used in the wars that NATO is waging around the globe, and this aspect is enough to shed light on the reasons why the US and NATO are trying to maintain their technological supremacy, while also using types of economic warfare against the BRICS and especially China.

The discussion so far seems to explain why the real question for the workers’ movement is not the dilemma of being technophobic or technophiles.

On the one hand, the problem is not technology, AI and robots. After all, they are nothing but results of social production itself, of accumulated human work experience codified as science, they are after all our own creations. All the more so since AI in particular encodes the very communication of millions or billions of people, i.e. it has a social character. Under the current circumstances, the conditions for the development and use of AI systems and new technologies in general are determined by the “masters of the earth”, the bourgeois classes and bourgeois states. Thus, there is no problem with new technologies as such. The problem is that today, and as long as capitalism exists, the exploitation of all these new technologies will be for the benefit of capital, it will be to ensure more profit and a larger rate of profit.

This course of development of the means of production and exploitation by capital means that new technologies, today will not be used to solve workers’ problems, but will objectively expand them.

• They will lead to deterioration in the working conditions of workers, to a reduction in employment and an increase in working hours.
• Their exploitation by the state makes the states more efficient in their operation, i.e. in heavy taxation of the people, in state repression and the surveillance of workers’ struggles.
• Their exploitation by imperialism produces new, even more dangerous weapons and leads to even greater power.

Thus, it is completely disorienting to think that the problems of the workers can be solved by new technology in the production process and in the country. New technologies are promoted with the criterion of increasing profits and the efficiency of the state, thus having a negative impact on workers. The forces, in politics and in the trade unions, which promote certain “technolutions” that put forward the position that more technology can be a basis for improving the lives of workers, are, consciously or unconsciously, cultivating illusions.

The real problem, then, is not whether we are deniers or whether we favour technology. Technology is not the problem. The problem is the economic relations in which it is developed and exploited.

We are positive that new technologies have the potential, in a radically different social and economic context, in the context of socialist relations, to play a very important role for the benefit of workers.

In a society free of profit and capitalist ownership, where production is developed to meet the expanded needs of the working people, new technologies, robotics, AI can be exploited with drastically positive results.

In production per se, new technologies, robotics, automation will be developed and exploited to substantially lessen human toil, to make tasks simpler for humans, and will lead to a rapid reduction of the working time, while improving the standard of living and the level of consumption.

At the same time, in terms of social organization, new technologies and artificial intelligence can offer major advances in scientific central planning, allowing problems that were previously much more difficult to solve to be tackled.

The possibilities of forecasting consumption can much better align production with needs, the possibilities of managing large volumes of multiplying data will greatly assist the scientific central planning of socialism to solve complex production problems.
New technologies will be used to assist education and health care services by increasing their efficiency and not as a cheap substitute for the popular strata as is the case today.

After all, the current enormous potential of the new means of production reflects nothing other than the same, ever deepening social division of labour, its social character. The new productive forces “cry out” for the need to change the relations of production, they cry out that capitalist relations are now more mature than ever. They cry out for the need for socialism.
Today, the workers’ movement must move forward using its economic theory as a beacon. It must move forward with a clear understanding of the effect of the means of production on labour and on the balance of power.

To realize that both sides that bourgeois propaganda highlights are equally ineffective.
Robots and AI are not “demons” to be exorcised, technology is not the problem.

At the same time, the exploitation of robots and AI by the state and capital will not solve our problems in the workplace, in society, instead it will multiply them. The advent of robots and AI does not make “the economy” better in general, it just benefits the capitalists to whom they belong.

Under the current circumstances, the international workers’ movement must fight and oppose the capitalist laws that seek to exploit the new technologies at the expense of the workers.

To elaborate a framework of demands that lights up the social reality, to explain that it is not technology but the capitalist owner that is the problem, to anticipate in good time the impact of new technologies in the various sectors and in their country. The need to drastically reduce working time while increasing workers’ wages and improving working conditions must be at the epicentre of their struggles. These struggles today will also pave the way for the great social changes that workers around the world need.

AI: Utopia or Dystopia?

Sudip Dutta (CITU-India)

The most widely debated emergence of Artificial Intelligence (AI) has brought into the famous question in the limelight – utopia or dystopia, which is going to be the final fate of humankind?

Will it be a world where labour will become insignificant or the material ground for socialist form of labour is peeping out from the germ?

In this discussion, we will try to see the AI from a political economic perspective rather than only a technological one.

So we will discuss its purpose, the alterations it is making in the instrument and methods of production, and effects on production relation, labour, economic and socio-political system.

The Purpose in general: Capitalism is a system dictated by the motive of limitless profit; in search for that, capital tries to expand the volume of sellable production to an unprecedented level. Machine was the starting point of Industrial Revolution installed that elevated the production to present gigantic stage.

In one hand, the labour of the workers is the only source of value for expansion of capital i.e. surplus, on the other hand, the organization of labour is the only deterrence in front of the unbridled pursuit of capital for surplus labour. Thus capital wants to get rid of unionized labour which can achieve heightened class consciousness and form a resistance, and simultaneously remains stuck in dire need of it in order to survive and thrive.

The dichotomy is expressed as peculiar relation of man and automation in capitalism, unforgettable friend and unbearable foe. Automation serves capitalism in 3 ways; by increasing the productivity of the whole industry, thus the volume of production and profit; by introducing new transient technology and thus enhancing productivity in selective monopoly houses, and securing higher rate of profits for them; and by repelling labour from job, thus enhancing the size of reserve army and imposing a downward pressure on wages.

So, productivity in general, productivity in particular and the exploitation capacity by increasing unemployment as a whole are the three gears of the automation vehicle.

Certainly, automation was very dear to capitalism and the sole ambition was the undisputed triumph of machine over man. To establish the autonomy of machines over human labour, capital replaced manpower by horsepower, wind, water and steam or electricity; the inanimate motive power started operating with or without the human muscle, but couldn’t get rid of the human intelligence which remained the sole overseer of planning, production, maintenance, marketing and all decision making processes.

AI and present state of Crisis: Capitalism is stuck into systemic crisis for last one and half decade. It has three established avenues to come out of any crisis (besides the commonly known unleashing of war); those are: the creation of the new branches of industry which are most profitable at that specific juncture of time; speeding up of the competitive productivity for the transient time, thus pushing over the burden of crisis over smaller companies and their workers; and the creation of massive unemployment, reduction of wages and thus transfer of the burden over the working class.

These three aspects are claimed to be very successfully met with the introduction of Artificial Intelligence (AI) and integrated Robotics. AI is supposedly the instrument to enhance productivity to a new scale and degree with downsizing various layers of employment. The crisis stricken capitalist camps are under extraordinary assumption that, through installation of AI, the productivity will increase without increasing cost over physical assets like machinery etc, followed by drastic economization of employment cost, and thus the profitability would rise, though less human labour will be engaged in a single unit of commodity.
What are the new instruments of production? A machine is an industrial robot if it can be computer programmed to perform physical, production-related tasks without the need of a human controller. The application of industrial robots includes assembling, dispensing, handling, processing and welding – all of these are prevalent in manufacturing industries; it also includes harvesting (in agriculture) and health inspection of equipment and structures (in power plants, transmission lines, petroleum refineries etc). Additionally, the ‘Internet of Things (IoT)’ are connecting machines and equipment to common network so that entire process can be operated remotely. In a nutshell, the entire revolution is about integrating the various manufacturing and service processes with a digital system and operate it remotely on the basis of decisions adopted through big data analysis.
The challenges were to develop software which can take and change decisions on its own and to develop software which can produce such software without human labour. And here comes the importance of AI. AI means machine that do not just carry out pre-programmed instructions but learn more new programmes and instructions by experiences in new situations and take automated decisions without a preceding codification. On the other hand, the software production was generally described as irreducibly communicative, ad-hoc and not amenable to Taylorist management method. In Taylorism, the scientific manager dissects the labour process, codify and optimize it.

The revolution in Machine learning is that, in a nascent form, it has developed the possibility of automating a process without a preceding codification. The procedure of machine learning generally operates in trial and error method at inhuman speeds, skipping over the codification of the labour process via sheer brute force. Simultaneously, the data science work that produces the software is in process of being dissected into components, fragmented, deskilled and automated with machine learning tools. The workers are automating their own work. This is certainly a revolutionary jump in the industrial scenario.

How it will change the work? The effect will be very wide and diverse in nature. In manufacturing, already the mechanism is established where a human works between two robots in a ruthlessly intensified labour process. Certainly, with increased power of remote surveillance and data analysis, the work processes will be more tightened and hectic thus weakening the bargaining capacity; more burden will be imposed upon the workers.

But, still now the technological advances are largely inadequate to develop AI robots to replace any simplest form of compound tasks performed by human. One Defence Research Test of Pentagon recently organized a competition where robots had taken an hour to complete a set of 8 tasks where as human did the same in less than 10 minutes.

So robots and AI are being engaged but the production processes are mostly remaining the same. Additionally, the human labour has been intensified, somewhere replaced. Man and machine, both are replicating each other in an unhealthy competition imposing unimaginable pressure on working class.

And there will be a peculiar change in the relation of production also! Rapid intrusion of automation of this latest kind needs a very fast adoption of state-of-art technology while continuously outmoding the necessity of dedicated-task specific experienced labour-power. The mode of production is moving towards a direction where the fresh trainee-batches employed most vulnerably are going to be burdened with the whole core responsibility of the production process. The Government schemes are also resonating the fact. Strikingly, AI will heavily impact upon the supervisory and monitoring jobs thus depleting workforce in those categories also. Thus, a material ground is emerging when employees of all strata, from trainees to executives will feel the need for getting organized to survive. Certainly, the digital integration of production processes will give enhanced holistic understanding to the new generation workers about the entire process of production and will heighten their consciousness.

Severe effects on labour: This deep automation will engulf all sectors, from manual to knowledge work. The lowest paid personal care jobs will increase while there will be decline in banking, production of goods and services, office and sales jobs. Goldman Sachs economists have predicted that the AI technology would bring “significant disruption” to the labour market, exposing 300 million full-time workers across the major economies; one-fifth jobs globally are exposed to risk of AI automation.

So, certainly there will be job loss in some sectors. But, will there be new jobs? Is the enhancement of Productivity is ensured? Professor Daron Acemoglu explains “if automation is “so-so” —it generates only minor productivity improvements — and creates all the displacement effects but little of the compensating benefits.” Indeed, the drive for extra profits from automation by leading companies can lower the general productivity growth. That’s because companies mainly introduce automation in areas that may boost profitability, like marketing, accounting or fossil fuel technology, but not raise productivity for the economy as whole or try to meet social needs. On the other hand, in the stagnation-inflation era of global economy, if the productivity grows, then also the demand driven output may grow in a slower rate, thus reducing the labour input and income (hours, employment and wages).

The latest ILO Report also ratified that the result of automation in the last 30 years has been rising inequality of incomes.
Finally, the sort of computer hardware systems needed to run AI costs huge, even before considering the data storage and networking. Undoubtedly, the introduction of AI investment will not lead to an easy cheapening of fixed assets and thus profitability in general will not rise automatically.

Though, there exists a good scope for monopoly tech-groups to enhance profitability by adopting this new technology but that will not be general or automatic. To ensure that, further downward pressure on labour will be enforced and unemployment will increase. But certainly, the supremacy of labour will prevail as labour is and will remain the sole agent of emergent creation and obviously, the only source of profit. The long conflict of labour with machine is going to reach at its highest stage when human will compete with its most advanced artificial matter, the AI.

Obviously AI is not a problem for a society where jobs and products are the right to all; the AI revolution reveals that we are entering in an era when human labour will be replaced by machines with unprecedented rapidity, where the scope for more creative contribution will unfold in front of humanity, an egalitarian era with more leisure and luxury of life, provided the reactionary capitalist social order is thrown off. The development of the productive force in this stage will not be any more compatible with this system. The coming days will be more challenging; new paths have to be traversed. It is the duty of the organized working class movement to convert this AI revolution into a weapon for the social transformation, not in a utopian but a scientific way.

LUEL (USA) Assistant National Secretary’s Comments

S.M. Cifone

First I would like to send solidarity from the class-oriented workers of the US to all the Palestinian and Syrian people throughout the world. My name is Scott and I am the Assistant National Secretary of Labor United Educational League in the United States. We are a movement of rank-and-file workers in the United States aimed and building a class-oriented center within the American trade union movement.

I would like to thank the WFTU, IWI for having this presentation today and I would like to thank the speakers today for a great presentation on the struggles the workers of the world face with the increasing implementation of AI. I would just like to give a quick update on some struggles facing the American working class due to the implementation of AI & automation. In October, the longshoremen on the East & Gulf Coast ports in the US went on strike due to the increased automation through AI instituted in the ports in the US. After 3-days on strike they came to a tentative agreement on wages and agreed to a 90 day extension on their expired Master Contract while they continued to negotiate on automation. Last month, negotiations were cut off and it appears they will be going back on strike January 15 with the goal of preventing automation on their ports.

Also, two years ago railroad workers in the US were blocked by the Biden regime from striking due to the Draconian Railway Labor Act signed into law here in 1926. Their main grievances were the railroads attempts to cut train crews to 1 person from an already reduced 2 person crew size due to automation. On top of this the railroads have introduced so-called “Precision Scheduled Railroading” which has led to cut inspection times & massive cuts in maintenance as well as expanded the train sizes which are as high as 3 miles long which lead to last year’s rail disaster in East Palestine, OH.

Labor United Educational League has started our research and are working on a plan to help the workers of America further understand the effects of increased automation and AI, as well as how to move forward as a class-oriented trade union movement. I must say, the presentations today have largely matched the discussions we have had on automation. One campaign are now working on is to struggle with American railroad workers to fight for the nationalization of American railroads in the Public Rail Now campaign. Under this plan the workers would hold a stake in the running of the railroads and could potentially mitigate the negative effects of automation and would allow its implementation to be done in the benefit of the workers and not the massive profits that currently go to the railroad cartels.

Thank You and Solidarity

 

See more photos below: