Have We Crossed the Rubicon Yet?

There was a joke --probably found in a Dilbert comic back in the 90s-- about how management felt that if they made everybody redundant the company could maximize their profits.

Back then, watching coworkers get replaced by overseas labor*, it certainly seemed like this was what Corporate America had in mind. The ultimate goal, however, was to use automation to replace the need for actual employees. Think of how robots moved into assembly lines back in the 80s, improving both quality and the need for extra workers, and you've got the idea.

Well, with the advent of Generative AI tools such as Chat GPT, Corporate America is going all in on using such tools to supplant actual workers. Unlike the robots of years past, this is aimed right at the white collar employees, but not just the ones in so-called "high cost" countries**, but all white collar employees. 

If the AI ingests buggy code, will it learn to write
buggy code? Blizzard is asking for a friend.
From LinkedIn.

If you told a C-Suite person that you could replace about half to 3/4 of your workforce with generative AI tools, they'd leap on it. "Imagine the cost savings!"

But those within the corporate world that are embracing Generative AI aren't thinking about their own jobs. Why would you need people in Finance if you had Generative AI to handle the budget and understood Finance Law? Why would you need people in bookkeeping if all you had to do was let AI manage receipts and pay requests, make judgements as to whether the request was legitimate, and make the proper payouts?

Or, why would you need people in a Legal capacity if AI had ingested all of the law in a locality and could make judgements based on both that and your company's contracts? Generative AI could even review case results and inform you which court to file a grievance in to maximize your chances at a favorable result. Generative AI could review all of the legal cases throughout a country (the US, for example) and determine how best to word your submissions to a court. And in which locale.

In a customer service job, an AI would be immune to social engineering attacks, as an AI customer service representative would stick to only what they are allowed to work with. Customers won't like it, but an AI being immune to any form of social engineering would likely balance out any public relations problems. 

For all of those gung-ho on AI in Corporate America who also manage by spreadsheet, I'm not so sure that they're aware that their own job can be replaced by Generative AI. All that has to be done is have someone prioritize the data ingested into the spreadsheets and just let the AI handle the rest. 

Imagine Bain Capital being just about a half dozen people at the top and Generative AI managing everything else, how would you tell the difference from the real thing right now? To the grunts or the people who buy the products that their businesses are selling, it would likely seem the same. 

Soulless investment firms aside, about the only thing that Generative AI can't do at the moment is sell something. Not the point-of-sale systems, mind you, but being an actual salesperson selling cars. The sales process itself is pretty well known, but people will likely prefer human interaction in a face-to-face environment over an AI. 

But who knows? Maybe that will change over time, but that also requires the risk of teaching an AI to dissemble or subtly lie, and I'm not so sure we want to cross the Rubicon there. Generative AI already "hallucinates" when it attempts to provide you with what it thinks you want, but intentionally lying for a separate agenda is a different thing entirely.

As for the end game, I suspect the sheer volume of investment dollars thrown at AI the past few years will eventually implode as such a bubble is unsustainable without a real return on investment. And let's be hones: a ROI I'm talking about is "I want to see profits next quarter" rather than "I want to see profits in 5-10 years". In a very real sense, Corporate America's short-sightedness may ultimately help employees in the short term, because investors expect to make a ton of money in return and they typically don't have the stomach to wait a decade for it to pay off. In the long run, however, most MBA work will be able to be run by Generative AI, so the same Finance Bros and Tech Bros who are all in on Generative AI may find themselves without a job themselves. Those same people had better start dreaming up what they want to do after they discover that Generative AI ate their lunch. 

*And then watching that first round of cheap labor from the 90s get replaced by even cheaper labor in cheaper locations in the 00s and then the 10s.

**Yes, that's what employees in the U.S., Western Europe, and Japan/Taiwan/South Korea are called in corporate-speak.