Groucho Marxism

The term ‘AI apocalypse’ is used to describe any of a number of hypothetical futures where artificial intelligence causes catastrophic harm to the human race. These futures can be broadly classified into four scenarios. The first involves mass automation replacing vast portions of the global workforce, leading to severe economic instability. The second involves large-scale automated cyber attacks targeting critical infrastructure like water and hospitals. The third involves the proliferation of autonomous weapons.And the fourth involves super intelligent machines developing goals that do not align with human survival or well-being. In this blog post I will go through the first of these scenarios and assess how worried we should be about it (I will leave the other scenarios for a future blog post).

AI is not developed enough to take over entire jobs right now, but it is already automating routine tasks and drastically reshaping how people work. Companies are increasingly using AI to handle repetitive tasks, leading to hiring freezes and fewer entry-level openings. This means that young people are being hit hardest by the AI boom and explains why there are so many NEETs: young people aged 16 to 24 who are not in education, employment, or training. The jobs most immediately at risk are customer support, entry-level coding, and administrative roles. Other jobs will be at risk in the medium-term as AI becomes more developed. Conversely, those least at risk are skilled trades, those involving human empathy, and those involving complex strategy.

In theory, AI should free humans up from routine tasks and allow us to focus on higher-level work. Indeed many commentators argue that this is precisely what will happen. Work published recently by Boston Consulting Group, for example, argues that AI will “reshape more jobs than it replaces”. Dig a bit deeper though and things don’t seem quite so rosy. Despite the optimistic headline, the research suggests that 15% of jobs in the US could be eliminated in the next five years, a pattern that will presumably be replicated across the Western world. Will sufficiently many jobs be created in the meantime to compensate for this decline? It seems unlikely to me. And that’s just in the next 5 years. It’s very difficult to predict what will happen beyond that, other than even more jobs being replaced.

It is clear therefore that in the not-too-distant future, AI will eliminate a significant proportion of jobs that are currently being done by human beings. So what are governments doing to prepare for this? Sadly the short answer is: very little. This is despite the fact that existing safety nets are woefully ill-equipped to cope with what is coming. Traditional unemployment insurance programs often require a specific work history, which can exclude new graduates and entry-level workers most vulnerable to early AI displacement. Unfortunately our politicians are asleep at the wheel and it is left to tech moguls like Bill Gates to sound the alarm. In a recent essay, Gates has warned that there is no plan for the upheaval AI will create.

So what should governments be doing to prepare? The two obvious solutions are a Universal Basic Income (UBI) and a Job Guarantee (JG). A UBI is a social welfare model where every citizen regularly receives a set, unconditional amount of cash from the government, regardless of their work status or personal wealth. A JG, on the other hand, is an economic policy proposal that aims to create full employment and by having the state promise to hire unemployed workers as an employer of last resort. There are arguments for and against both of these policy options which I won’t go into here (I will leave that to a future blog post too). But either would be better than the alternative: mass unemployment with no safety net. Unfortunately no politicians seem willing to even discuss these proposals at present.

Thus we find ourselves in the absurd situation where we are on the cusp of developing a technology that could completely free us from drudgery and allow us to focus on higher-level tasks; but instead of feeling excited about this, the overall feeling people have is one of fear and trepidation about the job losses that will result. If there is anything that demonstrates what a perverse system capitalism is this, it is this. Defenders of the status quo would argue that if it wasn’t for capitalism AI would not have been developed in the first place. That may or may not be true (personally I think it would have been developed sooner under socialism). But even if you accept this premise, you have to ask what the point of developing this technology is if it just ends up making peoples’ lives worse.