Another new peak hit in US stock markets last week, spurred on by signs from the Federal Reserve of anticipated further cuts in interest rates this year. However, it’s the ample cash reserves of leading tech companies and faith in their potential to monetise artificial intelligence (AI) that’s fueling the market’s enthusiasm.
However, we are being informed that AI will “alter the world”. It promises to drastically boost productivity (albeit at the cost of disrupting millions of jobs) and generate considerable wealth for global distribution. Also, according to an ARK Invest report last week, it projects a striking $40 trillion increase in global GDP by 2030 due to AI, as it will “revolutionise every sector, influence every business and stimulate every innovation platform”.
Despite this, the sheer jubilation and sense of certainty in this clear-cut story give me a feeling of uneasiness. Even for those of us who envision AI as the contemporary equivalent of electricity or the internet, we are simply at the inception of an intricately complex transformation spanning decades that’s far from conclusive. Yet, financial assessments are accounting for the complete paradigm shift. A report by Currency Research Associates in February noted that it would require 4,500 years for Nvidia’s future dividends to match its present price – a mind-bogglingly protracted timeline to consider.
Although Nvidia, unlike Pets.com, has solid revenues from tangible sales, the overarching AI narrative relies heavily on a multitude of uncertain factors. For instance, AI demands great quantities of water and energy. There is a drive in both America and the EU to encourage businesses to be transparent about their consumption. It’s plausible that these operational costs may skyrocket in future due to carbon trading or a levy on resource consumption.
Similarly, AI developers currently don’t need to possess the copyright for the content that trains their models. They also aren’t required to profit directly from AI; the expectation of future returns is enough to stoke the market frenzy. This undying techno-optimism and the mirage of inevitability are the Silicon Valley’s ways of accumulating perceived wealth. Though let’s not forget, a number of these “AI-focused” advocates were not too long ago promoting web3, cryptocurrency, the metaverse, and the benefits of gig economy.
People are starting to push back. The writers’ strikes in Hollywood were fundamentally about AI control, and unions are now addressing the broad issue of tech regulation.
Undoubtedly, AI has garnered notable validation from major, profit-laden enterprises like Google, Microsoft, and Amazon. However, skepticism exists amongst developers, even within these organisations. A top-level employee from a prestigious AI firm recently confided that the revenue predictions concerning the technology hinged more on conjecture than reality, confessing the presence of significant issues that need resolution.
This can be attested by those who have toyed with extensive language models. While researching for my projects, I wouldn’t count on a chatbot, as the accuracy of its data is questionable. I am unwilling to compromise on my capacity to filter my own informational inputs and would rather use Google for its clear display of sources and citations.
It’s imperative to note I belong to the high end of the white-collar job spectrum, but numerous uncertainties concerning how to infuse AI into workflows and its productivity versus human replacements exist, even for more monotonous mid-market tasks. There’s a bourgeoning backlash from humans, with Hollywood writers’ strikes primarily about AI control. Unions are also addressing tech regulation issues increasingly.
In addition, copyright retaliations against AI are escalating. Google faced a massive €250 million fine last week from French authorities for failing to inform news publishers about their articles being harnessed for AI algorithm training, and for not establishing fair terms for licenses. This action echoes previous lawsuits against Microsoft and OpenAI led by the New York Times. As AI delves into exclusive corporate datasets, copyright disputes are likely to multiply, perhaps even aligning with worker grievances over corporate intrusion.
Lastly, the control problem is a pressing concern. As Meredith Whittaker, the Signal Foundation’s president and AI Now Institute’s co-founder asserted in 2021, the core advances in contemporary AI are mostly the result of concentrated data and computation resources, managed by a few extensive tech organisations. Our increasing dependence on such types of AI, as per Whittaker’s assertion, hands over tremendous authority over our lives and entities to a select group of tech firms.
“The so-called ‘Fantastic Seven’ corporations have fueled enthusiasm for Artificial Intelligence (AI) and the surge in the stock market in the past year, causing the S&P 500 to reach an unprecedented level of concentration. However, as outlined in a recent report by Morgan Stanley Wealth Management, historically, ‘index concentration tends to self-correct,’ undermined by regulatory, market, and competitive dynamics, along with business cycle trends. The study further notes that ‘evidence implies equity returns have often found it hard to thrive following peaks in concentration.’
The mix of self-correcting elements could comprise the rise in major tech antitrust lawsuits. The prospect that carbon pricing and copyright penalties could pose challenges to the ‘free’ resources required for profit generation is also a factor.
Whether AI is perceived as the upcoming tulip mania or the next internal combustion engine, it’s worth scrutinising how the market values this narrative.
– Copyright The Financial Times Limited 2024.”