Why Some Say It’s Still too Expensive to Replace Humans with AI in Most Jobs

I’M NOT GOING TO spend a lot of time writing about the ebb and flow or the ups and downs of Artificial Intelligence.

There’s already a TON of that coming from everywhere else.

This isn’t a big surprise. As I wrote here recently,

“You’ll be seeing a lot more stories about the overhyping of AI in the weeks and months to come, but that’s to be expected because what goes up must come down.”

There’s a great deal being written about AI right now, and I know this because I share links to a lot of it here.

But out of all the AI-related stories that grabbed me this week, this one headline intrigued me the most. It simply said: AI is still too expensive to replace humans in most jobs.

It was in the weekly Build Mode newsletter from A.Team, a company that describes itself as “a members-only movement of highly-skilled product builders and bold companies teaming up to build what matters.”

Too pricey to replace humans?

HERE’S WHAT THEY SAID about AI still being too pricey to replace humans, and it is fairly short and based on research from MIT and the IBM Watson AI Lab:

“AI is still too expensive to replace humans in most jobs, a recent working paper from MIT and the IBM Watson AI Lab found.

As Michael Spencer and Tobias Mark Jensen explain in the AI Supremacy Substack:

“Previous studies on AI’s labor market impacts have assumed that any work task an AI system is capable of performing is exposed to automation without considering the technical feasibility or the economic viability of automating such tasks. This is quite an important detail. Just because a work task can be automated does not mean that it is desirable or feasible to do so from an economic and/or technical perspective.”

The researchers distilled this complex problem down to two questions:

1. 1. 1. Exposure: Is it possible to build an AI model to automate a particular task?
      2. Economically-attractive: Would it be more attractive to use an AI system for this task than to have human workers continue to do it?

More specifically, they mainly looked at the costs of fine-tuning and deploying a computer vision system to perform a task, including fixed costs, performance-dependent costs, and scale-dependent costs in a given time span vs. the marginal cost of compensation per worker.

They found that only 23% of visual-based tasks would be cost-effective for firms to automate. Put that in contrast to McKinsey’s prediction that activities accounting for 30% of hours worked in the U.S. could be automated by 2030.”

“There is some change that is coming”

There were a lot of other stories about the MIT/IBM Watson AI Lab paper on CNBC, Quartz, Computer World and Time magazine, among others.

As you might imagine, there’s a debate over how quickly and how deeply AI will impact the job market, and Time talked to Mark Thompson, the study’s lead author, who said this:

“Even though there is some change that is coming, there is also some time to adapt … it’s not going to happen so rapidly that everything is thrown into chaos right away.”

But even Time offers a challenge to that comment, adding this:

“A recent study from OpenAI estimated that 19% of U.S. workers could see 50% of their workplace tasks “impacted” by GPT-4 level systems — a far higher estimate than the MIT researchers’ study that focuses solely on computer vision. A crucial question for the economy in the age of AI will be whether the MIT study’s findings apply to more ‘general’ AI tools — ones that promise to automate most forms of cognitive labor that can be done behind a computer screen.”

HERE’S MY TAKE: There are lots of opinions about AI and its impact on the world of work.

Just reading a few of the many write-ups on the research-backed notion that AI is still too expensive to replace humans in most jobs will leave you shaking your head about what is REALLY going to happen with AI when it comes to people and jobs.

The short answer is easy: No one really knows.

“AI remains a less economical choice”

A LinkedIn newsletter that promises “fresh information for you about technology today” had this thought that resonated with me. It said:

“The journey of AI development is marked by a complex interplay of high costs and sector-specific challenges.

While the potential for AI to revolutionize industries like retail and health care is significant, its widespread adoption is hindered by the expenses related to technological upgrades, training, and energy consumption.

Currently, in sectors demanding nuanced human skills, AI remains a less economical choice compared to human labor in a swath of industry.”

Yes, how quickly AI will be implemented into the workplace will be determined on how much organizations invest in it.

That’s why the best story on all of this was an essay in The New York Times that focused more on what will happen whenever AI really STARTS to cut into jobs — When Your Technical Skills Are Eclipsed, Your Humanity Will Matter More Than Ever.

One last thought worth remembering

ONE THOUGHT RESONATED with me:

“Ultimately, for our society, this comes down to whether we believe in the potential of humans with as much conviction as we believe in the potential of AI.

If we do, it is entirely possible to build a world of work that not only is more human but also is a place where all people are valued for the unique skills they have, enabling us to deliver new levels of human achievement across so many areas that affect all of our lives.”

I’m going to remember those words moving ahead, because i really hope that when it comes to AI, they will really ring true.

 

Other trends and insights …

And if you need your weekly fix of AI news … 

• Business leaders using AI report higher job satisfaction, survey finds (From Tech.co/news)
• Managers, employees turn to ChatGPT to write performance reviews (From Axios.com)
• What Leaders Should Know About Measuring AI Project Value (From MITSloan Management Review)
• California plans to use AI to answer your tax questions (From SFStandard.com)
• Applicants say they are using AI tools in the job hunt (From HRDive.com)
• OpenAI CEO warns that ‘societal misalignments’ could make artificial intelligence dangerous (From APNews.com)
• Most employers lack guidance for using AI  (From HRDive.com)
• Is OpenAI’s text-to-video AI model Sora a blessing or a curse? (From CTech.com)

Dear Readers: I’ve been writing this wrap-up in one form or another for 20 years — from Workforce.com to TLNT.com to Fuel50 and now here. Let me know what you think at johnhollon@yahoo.com.