Data beats theory

Here’s a new empirical study of the impacts of automation and resulting employment changes on high and low wage employees over time.

Empirical studies are important because they replace conjecture with reality (but it’s easier to write a paper about conjecture.)

Automation’s long term effects on workers (5 years after a major automation event) are far larger than the short term effects, impacting both their economic and their overall well being.

Contrary to theoretical conjecture, highly educated and high-wage employees are displaced more often than their low-wage counterparts. But the latter are economically impacted for longer.

Caveats: The data are from only one country (the Netherlands) so more data and analysis is needed from other geographies and political climates as well. Further, the impact of automation on workers may change in the future.

AI returns for businesses have been elusive

MIT Technology Review is more often than not a booster of emerging technologies. A refreshing new article on their site is headlined

This is why AI has yet to reshape most businesses

The net of their analysis is no surprise: higher costs, greater difficulty, longer lead time and smaller gains. That’s no surprise to me.

I’ll add that the tools themselves operate at the giblet level (see: CEOs are seeking major business initiatives, not AI or other tech giblets.)

ERP, CRM, Supply Chain, HRMS and similar systems have been embraced because they’re not technologies, they are model-focused (domain specific) solutions, each designed with a model for how their core functions are supposed to operate for businesses and from a business point of view.

There have been major breakthroughs in AI technologies but, outside of what the big 12 AI firms (such as Google, Amazon and Baidu) are doing for themselves, there are no model focused AI based solutions. More on this in future posts.

Huge lags between technical innovation and GDP impact

MIT’s Erik Brynjolfsson, MIT PhD candidate Daniel Rock, and Chicago Booth’s Chad Syverson have published an AI economics brief worth reading. It’s a summary of a large body of work on why AI in particular isn’t boosting the economy yet, as evidenced by productivity and GDP data. Syverson’s productivity piece is particularly refreshing.

The Analyst Syndicate is now planning research on automation micro and macro economics. (This includes, but is not limited to, the effects of AI.)

Action: Share your specific economic-impacts interests and questions with us.

Disclaimer: This post is my own opinion. It’s not a paid post. I wrote it myself and I have no affiliation with any of the entities mentioned above.