10 June 2023 | PJ James, Countercurrents
The euphoria around many ‘frontier technologies’ such as Digitization, Internet of Things (IoT), Block-Chain, 5G, 3D Printing, Robotics, Gene Editing, Nanotechnology and the like that captured mainstream news headlines in the early decades of the 21st century has suddenly swept away by the recent rapid strides in Generative Artificial Intelligence (GAI). The discourse on AI that has been there for decades may be traced back to ELIZA (named after the fictional Eliza Doolittle from Bernard Shaw’s 1913 play Pygmalion) released by Joseph Weizenbaum in 1966, that allowed some kind of plausible conversation between humans and machines around a single keyword from users. The term Chatbot was coined in 1994 by Michael Mauldin referring to the online mimicking of human conversations through text or voice interactions. However, the emerging “super-intelligent” GAI chatbots that can reflect our tendencies and chat with humans using billions of words and oceans of data and information from the internet and fed beforehand into computer memory have no parallels in any previous experiments with AI. The new chatbots powered by Generative AI deal in algorithms or set of rules that are used to create new content such as audio, videos, code, images, texts and simulations. Chatbots with enormous information processing power at breakneck pace can analyze documents much faster than humans and are capable to summarise them within seconds.
Obviously, Silicon Valley-based Open AI’s ChatGPT4 (Chat Generative Pre-trained Transformer4) now being marketed as Microsoft-owned Bing ChatGPT, and Google’s Bard AI are the two leading corporate tech-giants today controlling the most powerful AI language models, though alternative versions of Generative AI such as Meta’s LLaMA, ChatSonic, Jasper, Open Collaborator, Wordtune, Claude, DuckAssist, Character AI, Replika, etc., are also there. Consequent on Microsoft’s launch of ChatGPT followed by Bing search challenging Google’s undisputed monopoly as search engine of the last two decades, the latter has resorted to its experimental Chatbot Bard leading to a technology war, as the inevitable outcome of cut-throat corporate competition. Since the very inception of ChatGPT in November 2022, the pace of advance of this chatbot has been 60 times faster than that of Facebook. Meanwhile, the US-NATO alliance has already expressed its deep concern on China’s AI program which, as reported, is proceeding at unprecedented speed. After banning ChatGPT (along with China, Russia, Iran, Cuba and Syria have banned ChatGPT), setting up broad goals up to 2030, China is engaged in transforming itself as the ‘prime mover’ in global AI development, even as NATO is pursuing for a deal with China over the use of AI in military. At the same time, proponents of ‘free software’ have come forward claiming that ‘Open Source’ will be the ultimate winner in the AI race.
Briefly put, the AI chatbot imitates human brain and thinking process and is capable of apparently doing what human intelligence can do. The Chatbot after seeking information from internet will give answers to our questions and for more information will even give the link to that page. While Open AI’s initial ChatGPT bot was trained on data that only covered up to 2021, the new Bing version covers far more recent events thereby enhancing its capacity to generate more efficient and accurate text. Chatbots can also give logical answers with reference to an uploaded picture on it. AI is also capable of performing a wide variety of tasks including voice assistance, recommending music, driving, health care and medical services, education, business, financial, legal and administrative services . Adaptive learning software and tablets created by AI can replace teachers and prepare question papers enabling students to go directly to the exam hall and even facilitating students cheat or commit ‘academic misconduct’. Chatbots can alter many spheres of human ingenuity and creativity including art and literature, may provide writing prompts for inspiration and even create articles, seminar papers, poems and songs, and in the process may mislead their users as if they are genuine and realistic. AI has the potential of spreading inaccurate and plagiarist content, political propaganda incorporating controversial viewpoints, responses and motives, often imbued with biased socio-political-perceptions, and culture distortions having even racist undertones in tune with the dominant societal consciousness. In the US, for instance, consequent on political parties’ use of AI to build out candidate biographies and generate video advertising, and based on the experience from proliferation of “deep fake” audios of opposition leaders on Social Media in the 2020 Election, a Bill aimed at preventing manipulation and deception of voters through AI generated political advertisements has already been introduced in the Congress.
On the Debate Around AI
Apart from the warnings of isolated scientists and technologists closely associated with the development of AI from the very beginning, of late, organized initiative from many quarters have come up against the ‘dangerous consequences’ of uncontrolled growth in AI technology. It ranges from the the Open Letter on March 29, 2023 by more than 1100 world’s leading AI experts, technologists and corporate CEOs including Apple co-founder Steve Wozniak and multi-billionaire Elon Musk to the latest warning issued by the Centre for AI Safety (CAIS) on May 30, 2023. The Open Letter has implored AI labs a 6-month pause on all AI experiments that are ”more powerful than GPT-4” till the development of safety protocols that can be audited by third parties and, asking governments, if needed, to impose a moratorium on such training systems. On the other hand, the San Francisco-based CAIS has issued a one-sentence statement thus: “Mitigating the risk of extinction from AI should be a global priority alongside other societal-scale risks such as pandemics and nuclear war.”
Those who uphold similar, strong but varying apprehensions include, among others, not only eminent persons like Geoffrey Hinton, the “godfather” of AI and Sam Altman, CEO of Open AI that pioneered ChatGPT, but also the late world-famous astrophysicist Stephen Hawking and even noted philosopher-historian Yuval Noah Harari. In fact, Stephen Hawking in 2014 itself had warned that “full artificial intelligence could spell the end of human race.” According to him humans who are limited by slow biological evolution would be superseded by AI which “would take off on its own , and redesign itself at an ever-increasing rate.” In November 2017, just 4 months before his demise, Hawking went to the extent of saying: “ I fear that AI may replace humans altogether. If people design computer viruses, someone will design AI that replicates itself. This will be a new form of life that will outperform humans.” At the same time, he acknowledged the great potential of AI research: “Because of the great potential of AI, it is important to research how to reap its benefits while avoiding potential pitfalls.” Yuval Harari, on the other hand, is of the strong opinion that AI has “hacked the operating system of human civilisation” and after “seizing the master-key of civilisation” is now leading to the “end of democracy” itself. Meanwhile Warren Buffett, US businessman-philanthropist has gone to the extent of comparing AI to “ the creation of atom bomb”.
These fears from the broad spectrum of concerned experts and technologists (majority of whom not for curtailing overall AI development) that emanate from the unpredictable and dangerous behaviour of AI systems and their suggestions for specific risk-control control steps such as ensuring transparency, appropriate regulation, supervision over software licensing, compulsory testing requirements for AI models above a certain threshold, etc., are not yet seriously addressed and even ignored by the corporate-tech giants having monopoly over AI Chatbots. As a manifestation their deep-seated antipathy towards people raising questions or doubting AI systems, in March 2023 itself, Microsoft had laid off its entire ‘ethics and society team’ with in the AI division, whereas Google fired some of its top ethical AI researchers for raising issues of bias in its algorithm. And in a blog post released during the third week of November 2003, Google has unequivocally put forward its opposition to “calls for a halt to technological advances”. As heated debates over regulation of AI are going on in US and EU( the latter has already drafted a Eurocentric AI Regulation Bill), India which is on the periphery of AI revolution, displayed little inclination towards regulation of AI. In a written parliamentary reply on April 5, 2023, IT minister Ashwini Vaishnav categorically said: “Government is not considering bringing a law or regulating the growth of artificial intelligence in the country. And Indian corporate big businesses being Silicon Valley’s junior partners have expressed their disinterest in bringing any control over AI. For, FICCI has openly come out with its position: “As AI comes in, less regulation is the best regulation.” However, following the meeting between Sam Altman with Indian prime minister Modi on June 8, the revealing sommersault came. Rajeev Chandrasekhar, the minister of state for Skill Development and Entrepreneurship reported on the government’s plan to bring a Digital India Bill to regulate AI. No wonder, whatever be the advances in science and technology, the laws of motion of corporate capital remains to be anti-people, and unless effectively resisted and overcome by world community, neoliberal capitalism shall continue to intensify its super-exploitation of working and oppressed peoples of the world, subjecting them to the biggest-ever AI-enabled surplus value extraction.
In this context, a glance at the historical background of the emergence of AI revolution that is destined to transform the entire labour process as well as human life would be in order. Of course, the potential impact of the unfolding trends in today’s AI and its application cannot be explained by merely comparing them with the developments during the epoch of Industrial Revolution or with the technological revolutions of the latter half of the 20th century or Web Revolution of the 21st century. For instance, the rapid advancements in production and processing technologies of the 1960s led to a multi-stage decomposition of production and its internationalisation on the one hand, and forcing down of value of labour power and super-exploitation of working class through an international division of labour, deindustrialisation, casualisation, and global outsourcing leading to the transformation of the whole world into a ‘wasteland of unemeployment’ on the other. Along with this, the material basis of neoliberalism that replaced ‘welfare capitalism’ was subjected to a structural transformation due to the large scale migration of capital from the productive sphere to the sphere of maddening financial speculation, where production is compared as a “bubble in a whirlpool of speculation” resulting in what is called a ‘post-industrial’ financialisation.
This was followed by the 21st century “digital revolution” and “digital relations of production” that are shaped by informal or unorganized, underpaid and super-exploited labour yielding corporate-finance capital inexhaustible and complex network of interconnected, global avenues of exploitation and plunder. Emergence of the transnational cyberspace as a tool of corporatization, communication and coordination enabled the corporate ruling classes everywhere to restructure the economic sphere and manipulate political and cultural systems by creating “intangible” or “virtual” data and “cultural products” or knowledge. This trend intensified further through the promptness on the part of corporate capital to seek further breakthroughs in labour-saving technologies including Silicon Valley’s initiative for AI development, diverting a portion of the trillions of dollars of public money gobbled up by Wall Street billionaires through what is called “quantitative easing” especially, following the world economic crisis of 2008. The outcome of this transformation is a far-right neoliberal and neofascist attack on world people on the one hand, and hitherto unknown levels of horrific wealth appropriation by a handful of big financiers on the other. Generative Artificial Intelligence is now becoming the most powerful tool for this corporate-fascist offensive in the hands of both big-tech billionaires and the regimes serving them the world over.
Tasks in Front
Today, the advent of AI revolution is taking place when the workers and oppressed peoples the world over are passing through the heavy burdens, direct and indirect, already superimposed by neoliberalism through such processes as deindustrialisation, financialisation and digitisation over the past decades. The new Chatbots propelled by generative AI though created by workers are now not only fragmenting, deskilling and denying them the knowledge required for production, and thus replacing humans with machines, but also taking over human minds and people’s thought-process too. This has enabled ruling classes and regimes to use AI as a tool against the ruled everywhere. No doubt, on the one hand, vast swaths of complex white-collar jobs will be taken over by machines, while on the other, labour will increasingly be used for “hybrid jobs” or “multiple jobs” which means workers will have to perform tasks previously done by multiple workers resulting in their extreme squeezing and intensified extraction of value. Many jobs pertaining to data entry, customer service, proof reading, translation, market research, social media management, telemarketing, news-reporting, teaching, para-legal and so on may cease to exist in the days ahead. According to World Economic Forum (WEF), the corporate-funded global lobbying think-tank, adoption of AI will result in a “new era of turbulence” impacting 25% of jobs over the next 5 years while Goldman Sachs has predicted a loss of 300 million jobs during the same period. At the same time, WEF has tried to whitewash the frightening situation with the claim that the money saved through this spectacular labour-saving move will be recycled back to the economy in the form of reduced commodity prices and increased purchasing powers prices of consumers.
In this context, first and foremost is the question of working class’ sustenance itself. Many analysts from varying perceptions have already brought out the gravity of the issue. In the first week of June 2023, IMF, the neocolonial-neoliberal arm of postwar imperialism has warned of AI-induced “substantial disruptions” in world job markets and advocated for governments to bolster “social safety nets” and work out appropriate tax policies rewarding companies to pursue a pro-labour approach. The past rounds of automation had affected factory jobs most, and based on the present pace of its advance, AI is estimated to impact on around 5 percent of the physical/manual workers in construction, maintenance, etc., in the immediate context. The explicit gender-bias of GAI is also very important, since the millions of office-based white-collar jobs that are going to be wiped out, majority of them will be female workforce. In many white-collar jobs like financial services while women comprise more than 50% of the office workers, their share in the senior management positions which are usually insulated from shocks caused by automation, women’s share is a paltry 2.5%. In addition to this, AI revolution will make millions of white-collar jobs obsolete. Apart from the 25% job loss over the next 5 years as already noted, according to Goldman Sachs, two-thirds of the current jobs in EU and US are also exposed to AI in the years ahead.
Amid the shock-waves sent by AI across the world, labour productivity is poised for the biggest-ever growth which under the existing relations of production will also result in spectacular wealth accumulation by corporate tech-barons. In this situation, rather than pursuing a “Luddite” approach which is counter-productive and negative, the immediate task is to find out an effective way of combating labour displacement, curtailing super-profit and increasing labour’s share in income. For this, the immediate demand of the hour is a drastic trimming of work-time, i.e, reduction in work to 3-4 days a week and effective resisting of the corporate-driven deskilling and fragmentation of the very same white-collar workers who created AI. Urgent interventions from working class and oppressed people together with all progressive-democratic forces are required to put forward a plan of preventing the diversion of the enormous productivity emanating from AI as super-profit of a few companies and for appropriate division of the technological gains for upgradation of the quality of life of the entire humankind. AI should be used in such a way as to entrust routine, repetitive, and boring functions with machines on the one hand, and increasing leisure-time and, thereby, enhancing human creativity and capabilities on the other. No doubt, the logic of corporate capital is not going to allow this smoothly. For instance, reports on Google using algorithm to fix who should be fired after tracking the movement of workers, the number of times they go to bathrooms, their time-duration for each bathroom-break, the number of times workers talk with coworkers, the number of norm-violations by them, and above all, to rate the productivity level of workers, and so on, are there. And, the process of AI-induced massive lay-off bloodbath has already been there in the agenda of companies.
It is obvious that corporate capital cannot think of a world without unemployment and super-exploitation of labour. The very design and purpose of the entire AI system developing today are solely motivated by corporate-super profit and, there can be no technological revolution that is class-neutral or external to the social relations under a regime of capital. Every advance in AI shall intensify the inherent contradiction between labour and capital leading to the displacement of growing number of workers from employment and super-exploitation of those who are lucky to be employed along with growing inequality and horrific wealth concentration. Sermons coming from neoliberal agencies like IMF and corporate think-tanks persuading big-tech companies and neo-fascist regimes who hold the reins of the system to manage the issues through tax measures and ‘freebies’ are also intended to divert world people’s attention from the systemic changes that are badly in need. However, such a political alternative being long-term in nature, all-out efforts for ensuring minimum means of survival are urgently required on the part of all those who are negatively impacted by the present developments.
Along with the direct survival issues of the alienated majority, the second threatening question is the environmental impact of developing, training and maintaining large GAI models which has been a less discussed one in mainstream media. While the declared motto of proponents of AI is to “make the world a better place”, and many AI experts claim that AI will help humanity battle climate change, a study entitled “Anatomy of an AI System” by Kate Crawford and Vladan Joler, published in 2018 much before the emergence of ChatGPT like energy-intensive GAI models has already demonstrated the unbelievable levels of energy consumption and carbon emissions associated with AI. Though it is difficult to exactly measure the energy costs associated with the complex AI system, many estimates have been there. For instance, the carbon footprint of training a single Big Language Model is equal to around 300,000 kg of carbon dioxide emissions– almost equal to the emissions by 125 round-trip flights between New York and Beijing. Today, a single data centre might consume an amount of electricity equivalent to 50000 homes, as reported in an MIT-related study. It is estimated that Google’s AI uses around 2.3 terawatt (one terawatt is equal to one trillion watts) hours of energy yearly. In view of the present pace of AI development at a global level, as reported, by 2025, around one-fifth of the available energy of Earth will go to keeping Data Centres in operation. To be precise, the rapid spurt in AI growth is worsening the environmental crisis, and has further sharpened the contradiction between corporate capital and nature along with the intensifying contradiction between capital and labour. Meanwhile, there is no dearth of usual corporate rhetoric, and as a manifestation, Google has declared its decision to run all its offices and data centres completely on carbon-free energy by 2030, that too through the deployment of AI itself !
And the third, but of utmost priority is that of combating the transformation of GAI as an all-embracing neofascist weapon today. For, unlike previous technological leaps that replaced physical skills of people, what makes AI revolution qualitatively different is its additional impact on the mental functions of humans which have its social, political and cultural ramifications. For instance, earlier, machines after receiving directions from humans were implementing them. If AI can transform towards making decisions by itself, then apparently, the line of demarcation between computers and human intelligence can vanish. If earlier machines could do only printing, now AI enables them to create texts and comment on them. However, the creation of texts, books and stories by machines/non-human intelligence not only makes them devoid of human ingenuity/craft but also leads to commodification of the latter. While AI has the potential to be deployed for helping humankind by inventing technical solutions, its mastery of language, capability of creating fresh data sources and working independently can create a situation where non-human intelligence formulating ‘stories’ altering history and resorting to ‘new history writing’, distorting culture and taking decisions impacting people’s consciousness, etc., which will be frightening. Rather than the outcome of human creativity and craft, book-writing by machines shall transform it into a commodity– commodification of culture.
This emerging situation gives those who are at the helm of affairs wielding political-economic power and hence having hold over AI unprecedented and untoward opportunities to manipulate public opinion in their favour. In a political context when the unholy alliance and merger between most corrupt corporate capital and far-right reactionary political forces are controlling the state regimes in many countries of the world, as already said, GAI can become the most effective tool for political propaganda, falsehood, and misinformation by neo-fascist forces, leading to an utter collapse of the so called bourgeois democracy itself. Its outcome will be a ‘post-truth’ world where freedom of speech and expression will be taken over by AI unless political mechanisms, countervailing checks and regulatory bodies are there to differentiate and distinguish between machines and humans. In the present neofascist situation where ‘official intelligence’ and artificial intelligence merge together or where chatbots speak on behalf of their godfathers and patrons, responsibility for evolving this corrective mechanism cannot be handed over to the regimes or corporate tech giants whose interests are already merged each other. Many proponents of AI have now started speaking on a ‘self-correcting mechanism’ (self-correction, introspection, conscience, etc., are human-specific) which can be inserted in to AI system. But the big question is whether it is possible, and if so, who can be entrusted with this critical task?
The neofascist dimensions associated with AI need not be nation-specific. As already noted, deployment of AI based surveillance system and harassment by government agencies and police directed against political opponents and human right activists based on plagiarist reports, fake photos, image impersonation, discrediting of opponents by planting information, which will be hard to detect can be routine things in a fascist situation. The same can be used by rival regimes against their opponents in their geopolitical tasks in other countries. Imperialist powers with their dominant position in AI can use it for advancing their political, economic and military agenda through superimposed treaties and agreements including even destabilizing political adversaries in remote continents. The latter is a global extension of the manner in which political campaigns against opponents are conducted based AI based analysis and research domestically. In the same vein AI can itself become an effective terrain for cyber-criminals too.
The upshot of the argument is that it is high time on the part of the people to intervene and block the flourishing of an AI regime that facilitates the continuation and further strengthening of the present inhuman system. This task cannot be entrusted with the existing regimes that serve corporate interests. In view of the warnings from AI experts that unregulated AI may lead to irreversible catastrophes, corporate centres and ruling regimes who are only concerned with their self-interests have come forward for regulation of GAI. The initiative on the part of British prime minister Rishi Sunak to host a Global Summit in London in this autumn aimed at devising international rules on AI is a typical move in this direction. This move prompted by the concerns from the development of “super-intelligent systems far exceeding human intelligence” and worries of AI being used to launch novel cyber attacks and develop devastating weapons, propagation of mass disinformation, etc., no doubt, is in conformity with the international interests of global corporatocracy and not based on the existential questions of the workers and oppressed peoples of the world bearing the twin the burden of corporate plunder and neofascsim. As reported, Sunak’s Summit intends to bring together heads of governments, CEOs of AI companies with the avowed aim of creating an AI watchdog similar to IAEA that monitors nuclear safety standards. As already noted, the aim of this Summit is to eliminate the risk of AI becoming a gravedigger of imperialism and to lay down the ‘rules of the game’ such as legal steps for avoiding intellectual property infringement by GAI.
At this critical juncture, it is up to the world people and all concerned to rise up to the occasion and come forward suggesting a political alternative to the threat emanating from unfolding AI regime especially with respect to the fundamental questions of livelihood, ecology and democracy as the immediate tasks, the details of which require further analysis, and not the purview of this article.
P J James is General Secretary of CPI ML(Red Star)