The amount of capital being poured into building out AI defies all historical precedent. AI “hyperscalers” are planning to spend over $5 trillion by the end of the 2020s—more than the total amount spent by all countries in World War II, adjusted for inflation. There has never been more money put toward a single objective in the history of our species.
The overwhelming majority of the financing—more than 70%—has come from private equity, as opposed to public markets and tech stocks. In order to entice investors to dump this much cash into their projects, top tech executives have had to make epic promises about what this technology will one day deliver. The result has been a mix of apocalyptic predictions and out-of-this-world levels of hype.
In the US, the prospect of attaining “Artificial General Intelligence” (AGI) has become the holy grail. We are approaching a horizon, the prophets of AGI tell us, where the technology will surpass human cognitive performance across all intellectual tasks. At that point, AI will be able to generalize, reason, adapt, self-improve, learn, and solve problems autonomously. When the race is won, whoever invested in the winning horse will have placed their bet on the most profitable enterprise imaginable.
No wonder the tech world is brimming with such unearthly optimism about the future. Sure, it could put a lot of people out of work, but just think of the trillionaires who will be minted in the process!
Meanwhile, in the real world, the workers who are being subjected to the reality of the AI rollout are decidedly less enthusiastic.
“Boring, bland, and lifeless”
So far, AI has had mixed success at replacing human labor outright. More than 600,000 tech jobs have disappeared since the rise of ChatGPT, but it’s unclear how much of this downsizing has resulted from adopting AI agents capable of performing the tasks that workers had been doing. Rather, many tech companies are conducting layoffs to prove their “fiscal responsibility” to shareholders at a time when the corporate debt binge is reaching unheard-of levels.
What is clear is that AI adoption has made work more grueling for human workers.
One sector where the progress of AI has arguably come the closest to replacing human labor is translation. Large language models trained on enormous amounts of text inputs have been the primary cutting edge of AI development. As a result, machine translation tools have come a long way since Google Translate first came out 20 years ago.
Even here, however, the job of human translators has been altered, not abolished. The task of rendering documents or generating subtitles for movies and TV into a second language is now done by machines. But those translations still need to be checked, corrected, and revised through a process called “machine translation post-editing” or MTPE.
The rate of pay for translators who used to charge by the word (for text) or by the minute (for subtitles) has fallen by 50% and 70% respectively. This forces translators to work longer hours, take far more jobs, and perform more cognitively demanding tasks for less pay. “It loses all the parts that I love about it,” one translator told the Financial Times. “It’s just a tedious job—boring and bland and lifeless.”
“AI seems to be having a de-skilling effect,” the FT reports. “Many experienced translators [are] refusing to do MTPE work, and as they leave the market, people with less skill and experience are stepping in.”
A growing number of companies are also deciding that machine translation is sufficient, dropping the MTPE layer altogether, and cutting costs by serving up sloppy, unnatural language as a final product.
An appendage of the machine
In May, the CEO of Standard Chartered bank informed shareholders that 8,000 employees would be let go in an effort to “replace lower-value human capital.” Workers were scandalized at being referred to in chattel-like terms, but the banker was merely speaking plainly about the real social relations of this system.
Whether or not the advance of AI leads to full-blown “jobspocalypse,” the capitalists are already seizing eagerly on its “de-skilling effect”—turning once creative and enjoyable tasks into mundane machine-tending.
From its earliest days, capitalism has used technology to squeeze more value out of workers in precisely this way. As explained in The Communist Manifesto:
Owing to the extensive use of machinery, and to the division of labor, the work of the proletarians has lost all individual character, and, consequently, all charm for the workman. He becomes an appendage of the machine, and it is only the most simple, most monotonous, and most easily acquired knack, that is required of him … In proportion, therefore, as the repulsiveness of the work increases, the wage decreases. Nay more, in proportion as the use of machinery and division of labor increases, in the same proportion the burden of toil also increases, whether by prolongation of the working hours, by the increase of the work exacted in a given time or by increased speed of machinery, etc.
Communists are by no means opposed to AI technology as such. We recognize its immense potential to make life easier—once it’s deployed in a conscious, rational way under the control of the working class. So long as it remains a capitalist technology, however, it will be an anti-worker weapon in the hands of our class enemy.