AI Is Coming For Our Jobs (And That’s Okay)

Humans are adaptable, and we have a long history of overcoming disruptive technologies.

By
47
Photo illustration featuring ChatGPT. (Photo by Nicolas Maeterlinck/Belg/AFP/Getty Images.)

Dear Capitolisters,

For a few minutes every Friday, I’ve been playing around with ChatGPT, the widely discussed “generative artificial intelligence” chatbot that can—after a few good human prompts—produce some pretty impressive writing about almost anything. If you haven’t tried ChatGPT yet, I highly recommend that you do; it’s easy and fun. My initial impressions of it and the other new AI products I’ve seen—for writing, art, coding, and other tasks—are that they’re still quite raw but clearly hold tremendous long-term potential for improving, if not revolutionizing, many aspects of our lives. And, judging by the explosion of investor interest in generative AI, including Microsoft’s massive bet on ChatGPT and the multitude of similar products coming down the pipe (here’s Google’s), I’m definitely not alone.

Accompanying this techno-optimism, however, has been a litany of concerns about the job-destroying potential of ChatGPT and other generative AIs. CBS News, for example, provides a handy list of the jobs “most likely to be replaced by chatbots like ChatGPT.” Others are worried about “mass unemployment,” “destabilizing white collar jobs,” or even a global AI “MegaThreat.” One recent columnist warned that “Artificial intelligence is not only coming for your job but will have a hand in laying you off, too.” And even more level-headed commenters, while recognizing the current technology’s limitations, warn that it will evolve rapidly and thus displace millions—just not today.

Tomorrow, on the other hand…

This content is available exclusively to Dispatch members
Try a membership for full access to every newsletter and all of The Dispatch. Support quality, fact-based journalism.
Start Your Free Trial
Already a paid member? Sign In
47
By
Comments (47)
Join The Dispatch to participate in the comments.
 
  1. Brook

    Another outstanding column. Thanks.

    Collapse
  2. Avatar photo
    Ben Connelly

    Great economic analysis and I especially loved the part about the Lump of Labor fallacy and the zero-sum mindset behind most economic fallacies (ie anything other than pro-capitalism positions).

    Still, I must disagree with you on AI. You’re probably right about the jobs. But AI must be destroyed. As a Butlerian Jihadist, I will not compromise on this position.

    Collapse
  3. Avatar photo
    Mick.Staton

    As someone who has worked the last 14 years in the software development field, I really have no concerns about AI coming for my job. The promise of what AI is supposed to be able to do is greatly exaggerated, and I expect will be the result of many lost fortunes, similar to what is happening in cryptocurrency today.

    As smart as people think computers are, a computer is the dumbest machine in the world. It can only operate on inputs and outputs, and it can only perform functions a human programmer tells it to perform, and even those functions are performed to the letter, without interpretation or intuition.

    What appears to be intelligence in things like the ChatGPT bot is really just a sophisticated searching algorithm with some semantic analysis and some pre-programmed behavior. The bot is not really writing a report for you. It is reading your input, using semantic analysis to determine the key search terms and parameters, then essentially doing a google search of internet content and returning the results in a text format.

    The bot did not research data and analyze it to consider its implications. It does not apply any intuitive reasoning to determine what facts should be included and what should not. It is essentially a summarization of content that already exists. The only really useful function this will serve is making high school and college reports easier (which will of course make our students dumber as they no longer actually have to learn anything).

    The danger AI like this ChatGPT bot represents is two-fold. First, as some people have found out, the bot has been pre-programmed with woke, liberal ideology, rendering balanced, unbiased reporting impossible. Second, since the only way this AI can respond is to summarize existing internet content, there is no filter on actual, verifiable information. This means that disinformation can and will be included in its reporting. This means that bad reporting, biased reporting, or simply stupid content, will ultimately skew the information this and future bots report.

    When you start adding content that is GENERATED by AI bots into the mix what you will end up seeing is a giant, worldwide game of telephone as bad facts and disinformation gets fed into AI that writes biased or incorrect content, which is then ingested by other AI bots that generate content that is even more skewed, and so on...and so on...

    As we say in my field, "garbage in, garbage out."

    Collapse
  4. Ck40579

    Your comparisons are myopic. When nail guns replaced hammers, more people became carpenters, and more houses per year were built. However, when the nail gun operates itself, what does the human have to switch to? The conundrum is that there simply are not new endeavours for humans to transition to, when more and more of any endeavour becomes automated.

    The transition from horses to tractors created a huge pool of labor, which were able to find usefullness by moving to cities and working in manufacturing. Nixon de-regulated and our manufacturing base was moved overseas, requiring a shift to a 'service' provision industry. Even if we could re-patriate that sector, the focus would be on automating as much of the process as possible. Too few jobs would be created in the requisite technical force to maintain robots.

    Today, outlets for words are scrambling to automate the process. More and more, algorithms produce content for text based media. The simple actuality of software being cheaper to purchase than keeping a human employed is immutable. Code to diagnose at the level of Primary Care Physician is not that far off. Code is already writing code. The sheer number of things for persons to do is not expanding, it is being narrowed.

    The only thing that could possibly change the outcome, is for collecting welfare to become a true career. The future is not bright, and far too many are wearing shades....

    Collapse
  5. Elderdxc

    "Last hired - first fired." When U.S. Steel automated many of its operations at the Gary Works plant, the workforce dropped precipitously. What was once a stable source of employment that helped drive the black middle class in Gary was now an absentee landlord, owning the lakefront but contributing almost nothing to the local economy or tax base. All because of, among other things, automation. There was no transformation of reshuffling for the black population of Gary, IN; there was either move or drop into poverty - or deal drugs.
    These transformations might work for the majority population because it will always look out for its own, but we will once again be forced to compete for the limited resources that remain, and we will be told, yet again, that we are just "playing the race card."

    Collapse
  6. Avatar photo
    Matt Anderson

    Some industries (e.g. law) are primarily self governing and likely have an incentive to maintain rules/codes of ethics which prioritize human interaction and human accountability. It'll be interesting to see how AI impacts these areas.

    Collapse
  7. MathewA

    Some good points here especially about the lump of labor fallacy. But aren't you also falling into a similar one? That the future will look like the past.

    Yes in the past we've had plenty of incredible changes. And yes people have always adapted. But AI combined with automation and robots really will be different. I can easily foresee a time in the not too distant future where robots and AI are simply way better than people at a huge amount of jobs. And that the vast majority of people simply can't outcompete robots. where the only jobs are those where people pay extra for "high class" human service (fancy resorts etc)

    Where society is broken up into two classes. The vast majority of people who don't do any work, but live on government UI, and a very small amount of people that own all the wealth and don't do any work either. And I guess another handful of people that make their money riling up the masses for revolution.

    Collapse
  8. Hoyasaxa2003

    I have never commented before on the Dispatch, despite being an early member, but this is a fantastic piece. Thanks for sharing.

    Collapse
    1. Avatar photocircle-angle-nib
      Scott Lincicome vertical-nib Author

      Thanks.

      Collapse
  9. Avatar photo
    Kurt

    Great piece, but I think it's wrong in a few ways. Now isn't like then. AI is different. What it all means is prognostication of the sort we don't have the tools for understanding.

    I'm not proposing some form of disaster, but I am proposing there's going to be a lot of people put out of work, and the horizon isn't showing us what might be coming to offset the situation. This isn't a "mechanical revolution" where new machinery (CNC for example) displaces older mechanical skills. This is different. Look at what it's done to the law or accounting professions. Most folks in those professions are only now seeing the picture, and the picture shows a lot of jobs that were once the introduction to a profession are now handled better by "AI". This will accelerate into all white collar work. Predicting otherwise is more hopeful than helpful.

    How it's different is still playing out and not well understood. I like the positivity of the piece, but I think it's philosophizing about new paradigms with old understandings and platitudes.

    Collapse
  10. Bskrantz

    Among the fallacies and biases people have when they think about work and jobs is a natural selfish perspective: what they would like to do and get paid for. Instead, people need to understand that work is a two-sided (at least) exchange of value. Perhaps we can prompt them to start recognizing this by asking them what they are willing to pay other people for (and how much they will pay).

    Collapse
    1. MathewA

      But what if the vast majority of humanity simply can't offer more value than AI or robot?

      Collapse
      1. Deleted Comment

        This comment has been deleted

      2. Avatar photo
        Thomas Grimmer

        In that case, there probably should be some manner of basic income.

        Collapse
        1. MathewA

          Agreed. I don't think we are there yet. But in 10 or 15 years I might change my mind.

          Although, there might still need to be some type of make work project. Idle hands and all that

          Collapse
Load More