Globally, wealth inequality is on the rise. Of the 10 richest people in the world (all worth more than $100 Billion), 8 are involved in the US technology industry. This includes CEOs of social media companies like Facebook and X, which extract value from our desire to socialize with one another in public spaces on the internet. As AI technology achieved a qualitative leap in its capabilities over the past few years, tech companies saw a new moneypot.

Despite the huge amount of money these billionaires have made, they have not used technology to make life better for average people. Cell phones get more expensive and less durable every year. Google Search has gradually become worse, burying useful information behind ads and launching an AI banner which confidently states false information. AI products largely are not useful investments in a better world, but ways for companies to pay fewer workers. It’s fine with Google if the product is worse, as long as costs go down. However, even on its own profit-seeking terms, AI’s success is overshadowed by a bubble of corporate grifters and gambling on speculative claims about future performance. Tech companies grew up in an era of extremely cheap loans after the 2008 financial crash. Companies did not use this money to invest in a safe and healthy future, but instead inserted themselves into our lives and became inescapable. Facebook renamed itself to Meta and has lost around $46.5 billion since 2019 pursuing virtual reality offices that no one asked for. Uber’s main innovation has been new ways to get around labor laws. 

AI Bubble

Currently, the mainstay of capital investments in technology is the AI bubble. Over $200 billion globally is invested in companies producing AI tools like large language models or voice assistants. This is driving massive investments in companies like the chip designer NVIDIA. NVIDIA made nearly $80 billion dollars last year, and its speculative value skyrocketed to $1 trillion based on the idea that AI will continue to grow exponentially, as will the need for new computer chips. 

Over the past few years, we’ve seen unbelievable claims around AI: that chatbots will replace teachers, doctors, and office workers; boost productivity; or even develop sentience. Now we can see the reality. We have AI tools that we didn’t ask for built into our phones and into Google Search. These are tools which can be sometimes helpful but routinely make errors and spout nonsense. AI tools used by companies haven’t raised economic productivity. In spite of the bubble, OpenAI is on track to lose $5 Billion this year. The main “success” has been deskilling workers like call center or helpline workers, or using a buggy AI tool to deny health insurance claims for the elderly. 

These aren’t growing pains. Data scientists have shown that general purpose large language models have reached a point of diminishing returns. The strength of a model like ChatGPT is its volume of training data and the current model is already trained on the entire internet and history of human writing. Better large language models may be possible for focused applications, but these require a lot of human labor, with experts curating a data set for a specialized purpose like medicine or science. AI companies aren’t interested in making focused investments for better tools like these, only in stealing your writing and art to use as training data.

This isn’t a new problem for capitalism. Every investment in automation gets more expensive, reduces costs by a smaller amount, and results in workers losing their jobs.

Inter-Imperialist Conflict

AI has also been in the news regarding the ongoing wars on Gaza and Ukraine. In Ukraine, drone warfare has seen record usage. In Israel, the army uses a computer system to identify so-called targets in Gaza. In reality, this system is finding increasingly tenuous reasons to kill civilians for any possible connections to Hamas, and mostly produces entire families as targets. In both cases, the threat is not some future hostile intelligence but the increased detachment and bypassing of the conscience of soldiers.

So, what is keeping AI going? We can look to US government policies like the CHIPS act, sending public money to support factories for computer chips and data centers in the US. The CHIPS act is part of the inter-imperialist conflict against China, with the US moving strategic production to more “friendly” locations that are better under US influence and control. This comes amidst global overproduction of computer chips, and AI projects provide artificial demand for these new chips. 

Environmental Impact 

Recently, the US sent $400 million to Taiwan’s GlobalWafers to build a semiconductor factory in Texas. Electricity is cheap in Texas’ deregulated electricity market and these factories use a lot of electricity. Microsoft is already the largest consumer of electricity in Texas through its data centers, largely for the computational power to run AI models like ChatGPT. The data centers and computer chip factories planned in the US will double to 35 gigawatts by 2030, consuming nearly half of all planned renewable energy projects and offsetting our paltry progress on fighting climate change.

These investments prioritize companies and leave regular citizens with crumbling infrastructure. The deregulated electricity grid in Texas is in shambles, and extreme weather leaves millions without power for days. Extreme cold weather in 2021 caused weeks long blackouts which left 246 dead and 12 million people with burst pipes or other major expenses. This year, a hurricane left 2.6 million people without power for 10 days during a heat wave, and 23 people died as a result. During these outages, corporate clients like Microsoft were given priority over the lives of working people, and did not lose power. These extreme weather events will only become more common as climate change continues unaddressed. 

We cannot rely on capitalism to solve its own problems. Energy investment is weighted 95% towards making oil extraction more profitable, prolonging the lifetime of fossil fuel infrastructure. Capitalists have spent money to build pipelines, drilling rigs, energy-hogging data centers and other expensive infrastructure without a long-term vision. Wind power costs as low as 1/10th to produce compared to fossil power, but the fossil company infrastructure exists, and at this point amounts to free money for its investors – they will not give that up without a fight. Climate change is an urgent problem that requires action – it cannot wait 100 years until we run out of oil to begin solving it. 

Break From Capitalism

Tools like AI could actually be made more useful and save us time doing undesirable work. If working people had a say, we could invest in automation that makes our jobs easier without worrying about profit. Workers could have a sliding scale of hours that means if your job takes less time, then you get more time back with no loss in pay. Autoworkers in the UAW fought for a 4-day workweek in their strike last year, as the building of electric vehicles takes fewer hours than building a gasoline powered car. They didn’t win that, but it showed what working people should demand in the era of increased automation.

Real innovation requires the ability for working people to make changes from the bottom up and wield their creativity and expertise. We need to bring technology and energy companies into democratic public ownership. This means working people would have direct control over how these companies run, to ensure they function for the public good rather than short-term profit seeking. What will fix climate change is investment in our basic infrastructure and prioritizing our resources in a way that benefits regular people, not massive corporations. New technology can be used to improve our lives and reduce our workloads, but only if workers have control over what investments are made and how these technologies are used.