Artificial Intelligence, With Problems Galore, is Getting Boxed In

Five weeks ago, I published a post about an artificial intelligence backlash, or reactions against it from various directions.  Since then, when we factor out the news stories speculating about AI’s effects on employment, telling us about new or not-so-new robotics advances, and subjective investment advice, there hasn’t been much good, and it adds up to something worse.

First, as shown in “CheatGPT,” by Aki Ito in Business Insider on August 1^st, we have workers using AI against company rules, taking advantage of lagging or overly cautious policies as well as thoughtfully reasoned ones.  That is really an issue between employer and employee but could turn out to cause the former real problems when accuracy is not what it seems.  Such isn’t rare, as “When Hackers Descended to Test A.I., They Found Flaws Aplenty” (Sarah Kessler and Tiffany Hsu, The New York Times, August 16^th).  At the Defcon hacker’s conference this month, participants “tried to break through the safeguards of various A.I. programs in an effort to identify their vulnerabilities.”  They were able to get the software to make racist and discriminatory statements, select sample job applicants by Indian caste, and write factually about nonexistent people, places, and constitutional amendments.  Efforts like these, in this case strictly benevolent, are invaluable, with experts revealing previously unknown flaws – the bad news is that those shortcomings are real.

Is, or will it, be true that “AI is ruining the internet” (Arantza Pena Popo, Business Insider, August 8^th)?  The programs are now able to defeat many of the human-verifying “captchas,” causing sites to make theirs so difficult many actual people cannot get through.  A Europol study estimated that within “a few years,” ninety percent “of internet content” will be produced by AI systems and will contain significant errors.  That could result in online resources seeming “engineered for the machines and by the machines,” its value could greatly decrease, and it may in turn poison future AI improvements.

The most telling of these articles, though, is “Digging for digits” (The Economist, August 19^th).  It told us about how the world is running out of accessible, non-copyrighted data, for which companies are competing and paying increasing amounts to use.  Some of that will be inputs into ChatGPT 5, which may be released before the end of the year.

But how about ChatGPT 6?  If we see that one someday, it may not be fundamentally better.  Why?

Although conceptually it could be boundless, artificial intelligence is being stopped on several different fronts.  First is copyright problems, with unauthorized incorporation of legally protected work well integrated into ChatGPT 4, and certain to result in further lawsuits and massive settlements, with a chance that its entire knowledge set could need to be erased and rebuilt.  Second is chip shortages – developing large language models at contemporary strengths consumes more silicon than it can get.  Third is calls for more regulation, which, with governmental authorities perennially behind the curve with technical issues, could impede or even halt development, production, or both.  Fourth is an upcoming data shortage as above. 

The fifth limit, and perhaps the ultimate, is money.  ChatGPT 5 will cost in the billions of dollars.  Since growth is exponential, there is a real chance that ChatGPT 6 or its equivalent would run in the hundreds of billions.  Even in as future-oriented a field as information technology, companies, venture capitalists, and other lenders and resource providers want results quickly and will see the other four problems. 

Six years ago, driverless cars seemed poised to take over world roads within a generation.  Now, although they have done well in useful if controversial niches such as taxicab travel, and their technology is turning up more and more in human-driven autos, their overall development has apparently stalled, with little press about them becoming the norm.  A serious around-1970 prediction said that within two decades artificial hearts would be, after cars, the second largest American industry – it is now possible to get one, but its implementations are measured in dozens.

The same may happen with artificial intelligence.  It may find its role in business writing, and as a lower-end substitute for human effort.  By 2040, people may laugh about the idea that this business tool could take over the world, and compare its dangers to those of copiers.  Or not.  Will artificial intelligence stay way short of its potential?  I don’t know, but we need to seriously consider that possibility.

Uber, Lyft, and Gigs in General – How Are They Doing?

A lot has piled up here about working without conventional jobs.  Let’s see what it is.

Oldest is “Biden Proposal Could Lead to Employee Status for Gig Workers” (Noam Scheiber, The New York Times, October 11^th).  That wouldn’t be wide open, but rather “lowers the bar for that employee classification,” by incorporating “how much control workers have over how they do their jobs and how much opportunity they have to increase their earnings by doing things like offering new services.”  The piece, though, is more about how current laws are often disobeyed by companies preferring workers to be contractors.  Consistently correct categorization would do more for those with gigs than changing it.  As for Megan Henney’s October 12^th Fox Business response that “Biden’s proposed gig worker rule could deal a major blow to small businesses,” the end of slavery did that as well.

Also on the 11^th, The New York Times told us to “Meet the App That Helps Gig Workers Know How Much They Really Make” (Erin Griffith).  It’s Para.  Not a cost-accounting product, but one that gets into company databases and reveals how much workers will be paid, often including tip amounts, so they can accept or refuse giving rides or delivering restaurant orders.  The product delighted neither Uber nor DoorDash, who both sent Para “cease-and-desist letters.”  Para provides information important enough for one driver to report that “a no-tip order from Red Lobster could mean less than $3 for 30 minutes of work during a peak time.”  Overall, if companies freely provided this information, there would be no need for anyone to get it any other way.

From the front ridesharing lines came “What Shocked Me Most When I Became a Lyft Driver for a Week” (Timothy B. Lee, Slate.com, December 22^nd).  The reporter found great variation in his portion of what Lyft charged, from “less than 30 percent” to “more than 80 percent,” and over the week collected 52%.  Interesting, since the rate Yellow Cab of Milwaukee paid when I worked there in 1976, for those preferring a percentage of fares, was 55% – with the company providing the taxi.  As Lee put it, “it seems that the smartphone revolution didn’t make taxi dispatching cheaper and more efficient.  It made it way more expensive.”  So, perhaps, could someone create a cheaper rideshare model?

Three articles on the other large de facto cab company appeared in the New York Times over five weeks.  “Uber Drivers Say They Are Struggling: ‘This Is Not Sustainable’” (Winnie Hu and Ana Ley, January 12^th).  That’s no wonder, if they need to pay for their cars out of 52% of fair market fares.  Yet they have learned something – one interviewee said he was leaving to drive a truck.  Also, little shock that “Uber Reports Record Revenue as It Defies the Economic Downturn” (Kellen Browning, February 8^th).  It was $8.6 billion from October to December 2022, of which $595 million was profit.  Yet “Financial Woes Thrust Lyft, Long in Uber’s Shadow, Into the Spotlight” (also Kellen Browning, February 15^th).  On their “record revenue of $1.2 billion in its most recent quarter,” they lost $588 million.  Strange.

Moving almost six months later, we heard that “Minneapolis Uber, Lyft drivers want minimum wage, companies say it could be worse for riders” (Mills Hayes, Fox Business, August 2^nd).  I could not have imagined that as a cabdriver, when, over full-length days, I netted over double the $1.80 per hour I got elsewhere the next year.  And yes, Uber and Lyft opposed that, perhaps because it illuminated how little many earn after expenses.

Last, could it be that emphasis on gigs is causing “The death of hobbies” (Eve Upton-Clark, Insider, July 31^st)?  Not really.  Since my early teens, I have consistently sought out ways of earning money from what I enjoy, and have done that, with different things, most of my life.  That does lead to wanting a true avocation when others become vocations.  Yes, as the author says, going for profit can cut down other activities in that area, as, for example, reselling collectibles cuts compatibility with keeping them for yourself, but people can always adjust what they do.  A great deal of the fun remains – and the advantage of earning money instead of spending it can, over decades, be remarkable.  So, find the right choices for you and go right ahead – you’ll probably like it.

Hiring Policy – What’s New and What Could Be Next?

Something not getting much attention, except recently with artificial intelligence making inroads and causing controversy, is how employers hire.  These articles are spread over ten months, but it takes a while to get news on this subject, about which companies can be secretive.

Finally, “Globalization has come for US tech jobs” (Aki Ito, Business Insider, September 28^th).  What took it so long?  I predicted in 2012’s Work’s New Age that that would happen quickly, given that most such positions need little or no in-person presence, and Russia and India, to name only two, had many skilled candidates not expecting American pay scales.  But now we can read about “the payroll startup Deel,” which “wants to prioritize hiring the best candidates regardless of location,” and is looking on at least three continents.  Per this piece, companies learned how people can collaborate remotely from the pandemic, so some are taking that further.  Dare I say that we should expect much more?

After that is “5 jobs U.S. employers plan to outsource,” by Lee Hafner on November 17^th in Benefit News.  Really fields instead of positions, they are human resources, information technology, marketing, sales, and software development and testing.  That’s two in computers, a departure from IT’s frequent appearances, as late as last year, in listings of best careers.  The others are somewhat surprising, as some sales, HR, and marketing research must indeed be done in person.  It is possible that this list, after the remote-work backlash, would be different now.

Something worth considering is Brock Dumas’s December 23^rd Fox Business “Trouble filling a job?  Look at hiring someone with a criminal record, HR pro says.”  This article and the next address re-examining hiring restrictions, appropriate when potential employees are scarce.  The author cited studies showing that employees with such history were not only as reliable and effective as those without it, but often more, as they have “fewer options.”  As well, “former offenders” can be “less expensive to hire.”  People can be given positions less pertinent to their previous crimes, and their status may not be relevant at all.

Another approach, taking place in other areas, is to evaluate credential requirements, many of which were installed when hirers wanted to thin jobseekers’ ranks.  The largest is the subject of “See Workers as Workers, Not as a College Credential” (Editorial Board, The New York Times, January 28^th).  Points made here were to follow the example of Pennsylvania governor Josh Shapiro, who, ten days before this piece’s press time, “eliminated the requirement of a four-year college degree for the vast majority of jobs in the state government.”  It is easy to use graduation as a proxy for various general skills and attributes, but interviews, among other tools, can serve that function also.  We may not go back to the age, during some of our lifetimes, when people wanting business careers usually did not attend college, but we can loosen general education’s grip.

We may forget that, just before the big artificial intelligence news, 100,000 people in information technology lost their jobs.  Per Tyler Le in Business Insider on February 14^th, “Big Tech’s massive layoffs will come back to haunt it.”  The author maintained that the bloodletting has badly hurt the field’s reputation, especially for a generation valuing stability, so IT may no longer be “the destination of choice for the smartest kids at the best schools.”  But what goes down often goes back up.

I finish with a newer hiring-related issue, as “Affirmative Action Ruling May Upend Hiring Policies, Too” (Noam Scheiber, The New York Times, June 30^th).  “The Supreme Court’s rejection of race-conscious admissions in higher education” has been rippling outward, with potential for new attitudes even in legally unrelated areas.  With the lawfulness of hiring policies favoring certain groups already in doubt, we can look for many changes as new practices take hold.  As with the other issues raised here, that may take a while.

This Morning’s Jobs Report Reflects a Peaceful and Prosperous July, with AJSN Showing 88,000 Lower Latent Demand

The new Bureau of Labor Statistics Employment Situation Summary was not expected to show any sudden changes, and it didn’t. 

I saw estimates of 180,000 and 200,000 net new nonfarm payroll positions, straddling the 187,000 outcome.  The seasonally adjusted unemployment rate shed a tenth of a percentage point to reach 3.5%, and the unadjusted variety held at 3.8%.  The count of jobless fell 200,000 to 5.8 million, with that of those with jobs up 433,000 to 161,982,000.  On the heels of two one-million-range drops, the number claiming no interest in work took a breather, gaining 25,000 to 93,070,000.  Long-term unemployed, or those jobless for 27 weeks or longer, reversed its 100,000 June improvement, returning to 1.2 million, but the number working part-time for economic reasons, or looking unsuccessfully for full-time work while keeping one or more shorter-hours propositions, dropped 200,000 to 4.0 million.  The two measures showing how many Americans have jobs or are one step away, the labor force participation rate and the employment-population ratio, held at 62.6% and gained 0.1% to 60.4% respectively.  Average private nonfarm payroll hourly earnings added 16 cents, well over inflation, and are now $33.74.

The American Job Shortage Number or AJSN, the statistic showing how many currently unavailable positions could be quickly filled if all knew they would be easy and routine to get, came in as follows:

The largest AJSN change from June came from those wanting work but not looking for it for the past 12 months, and that reduced the outcome by only 111,000.  A rise in those claiming discouragement and a fall in those wanting work but currently unavailable for it were the only inputs adding or subtracting more than 18,000.  The share of the AJSN from those officially unemployed increased 0.3% to 34.4%.

Compared with a year before, the AJSN has lost 379,000, over 80% from a reduced count of those wanting a position but not looking for it for 12 months. 

So – did anything happen here?  Yes, although the changes were small, we continued on the right track, including people rejoining the labor force.  We are still adding more jobs than our population increases can absorb.  Although the count of those saying they do not want to work went up, that was a trivial amount when compared with the number of Americans reaching 60, 65, 70, or any other reasonable retirement age.  Our national employment status is vibrant and robust.  Once more, the turtle took a solid step forward.