Is AI coming for my job? Not yet, but it will, by stealth
Anthropic release a report on "Labor market impacts of AI: A new measure and early evidence" this week.
It looks at a measure it calls Observed Exposure. What is AI is actually impacting, rather than predictions of what might happen.
The report notes that:
1. There is a large gap between the capability of AI, and the current deployment
2. Impact is not from just job losses, but reduced hiring into roles
3. There is little evidence that AI is driving unemployment figures
4. Hiring of younger workers in exposed roles is slower.
Is it a case of hype over reality? That the tens of thousands of job losses, allegedly due to AI have been quietly absorbed, or that they are not significant enough to impact the totals?
Is it a case of 'not yet'? It is clear that the AI usage today covers a fraction of what is theoretically possible. That the displacement will increase with the adoption of AI within organisations. It is possible.
What are the signs? You need to look at both ends. Jobs being displaced, and lower hiring rates. It makes it harder if you are entering the workforce. Unlike Covid, which had a dramatic impact, this is more like to be a slow burn.
Is it accurate? It is too early to tell. they are laying down the measure as something more meaningful to measure over the coming years.
Who is most exposed: Software programmers, Customer service representatives, Data entry clerks, Medical Record Specialists, Market research and marketing analysts, Sales representatives who solicit orders, Financial and Investment Analysts, Software QA and testers.
The report suggest that those are most exposed (top quartile) are female (54%), white (65%), married (55%), with a batchelors or graduate degree (55%), and an hourly wage of $33.
SOURCE
https://www.anthropic.com/research/labor-market-impacts
BE SCI AI OPINION
I can't help feel that this report needs a pinch of salt.
With the criticism being laid at the doors of the AI Frontier model companies that they are displacing jobs for their own benefit, this feels like an argument to say 'we haven't had a noticable difference' and that it is 'hype over reality'.
But is it?
There is a good argument that organisations who have been bloated for a while, thanks to their high margins and an inexperience of reducing headcount have used AI as an excuse for a while - who can forget the CEO of Intuit when he fired a chunk of his staff, ahead of the curve becasue they were the wrong kind of people.
It is no surprise to many organisations that the impact has not been fully felt yet. When (if?) the potential is realised then maybe the situation will look different.
It argues that the big move to 'offshoring' was also seen to remove significant jobs, yet the employment levels did not fall as much, or were balanced by other roles. It does not speak to the quality of work left behind, just the quantity of unemployment.
It feels as though this will be a 'boil the frog' strategy, where the gradual demise of roles, from both ends (hiring and firing) will creep up on us.
