Are our (White-collar) jobs doomed with AI?
Spoiler alter, I believe the answer is “Probably, yes”.
I confess I am one of the Cassandras when it comes to the impact of AI on (especially, but not limited to) white-collar jobs. However, I believe we are living in a pivotal time in human history; much like the advent of fire, that of the wheel, or the steam machines heralding the Industrial Revolution and the new opportunities they brought. If used ethically and equitably, AI has the potential to solve many of humankind issues (e.g., diseases) without creating too many downsides (e.g., mass surveillance or increased imbalance wealth distribution).
I base my hypothesis on three observations:
First, corporate value creation:
No company will refrain from any opportunity to increase its productivity (and hence profitability) within the legal framework of its activities. In other words, if a new tool or process allows a business to foresee bigger returns, it will inevitably seize the opportunity. This is, in fact, the true raison d’être of any ongoing concern. The AI Revolution is a godsend opportunity for business leaders to create more value, more efficiently for shareholders.
Second, regulatory oversight:
Governments and regulators are not going to get in the way of the AI Revolution and impair its potential. Think about it: what is there to gain for a leader by getting in the way of progress? Especially when there is a struggle happening at the nation vs. nation level for AI dominance, with geopolitical implications.
For example, the United States, with its rich history of technological innovation and a robust ecosystem of AI research and development, has been at the forefront of AI advancements. Meanwhile, China has rapidly emerged as a formidable contender, more than any other country, in the AI race: the Chinese government has outlined ambitious plans to become the global leader in AI by 2030. This competition between the two nations is characterised by several key factors: talent acquisition and retention, data access and utilisation, technological breakthroughs, commercial applications, and regulation. Regulation will be used to accelerate AI development, not stall it.
A good proxy for my intuition about AI regulation is the content of the GDPR. Although the intent of this framework is to refrain companies from doing “bad things” with your data, some room has been left intentionally for business to acquire and process it. There are certain aspects and circumstances where some degrees of flexibility or exemptions exist. For example, ‘legitimate interests’ allow organisations to process personal data without explicit consent if they have a ‘legitimate interest’ that is not overridden by the individual’s rights and interests. You need to go deep in the consent form to explicitly deny a company willing to capture your data the chance to do so. Lawmakers will not want to be the ones stifling the potential of their national tech industry.
With the final version of the AI Act expected for 2024, we should not be surprised to see such opportunities for companies to process our data as they see fit.
Third, Productive efficiency and market equilibrium: In other terms, the market favours the most efficient production factor and in an efficient market supply and demand tend to match each other.
The industrial revolution has brought about more needs and, as a result, curtailed the then-dreaded mass reduction in the need for human labor compared to that of machines. This phenomenon was due to the fact that the new market demand outmatched (and still does) the supply capacity of industries. For example, the car industry is heavily automated, yet there are still places where people are eagerly waiting to rise to the middle class and be able to afford a car. In other words, the Industrial Revolution, instead of flooding the market with all the products we needed to produce, has uncovered new needs that it has not been able to completely satisfy. On the other hand, the AI Revolution has created the potential for companies to exponentially increase their output. For instance, developers using AI copilots report an increase in their overall productivity. However, our demand for products and services typically provided by white-collar workers has not increased significantly. Unless we discover an entirely new type of social media, engineers who have spent the last ten years tinkering with pixels on your favorite social media platform to keep you hooked will be replaced by an automated process. In a nutshell, AI has the potential to exponentially increase supply, while the demand is stalling, leading to a self-regulation of demand by eliminating less efficient production factors from the equation. Can you guess who that might be?
TLDR;
Obviously, I do not believe that all white-collar jobs will disappear; they will undergo ‘reconfiguration’ in terms of quantity and content to align with market needs. Whether business ambitions or regulations will influence this transformation is not essential in understanding the future of human labor. The crucial factor will be the equilibrium between supply, referring to the quantity of services competitively offered by human labor, and demand, representing the required amount of services. Unlike manufacturing, where producing a new item entails the entire production process (and costs roughly the same as the previous item), AI services can be replicated at a marginal cost.
