There are many uncertainties about the impact of Artificial Intelligence (AI) on the future of society. Despite the current AI frenzy, by all accounts, we are still very early in this transition. But one outcome that I was recently introduced to, and which has gotten less media coverage is that AI threatens to create a tremendous overhang of what’s called “technical debt,” or usually just “tech debt,” with potentially few people who will be able to address it.
“Tech debt” does not mean borrowing financial assets, but rather it is a specific tech industry concept. It works like regular debt: you “borrow” by taking a shortcut that gets you to your goal faster at the cost of something you must “pay back” later. You pay it back by replacing the shortcut with a robust solution when you can. When software developers take shortcuts instead of building clean code, they burden their company with future extra work and cost. Excellent coders used to be at a premium for creating elegant code that avoided this. But now, as anyone can be a software developer using AI coding tools, companies are creating programs that generate a desired outcome but often are rife with hidden inefficiencies that will only show up when future updates or changes are requested. This inadvertently creates invisible tech debt. Anyone using AI knows that while it may be a quick and easy productivity gain, the work itself is very often of a low quality and sometimes it is hard to determine how it was generated. With so many people now using these tools but lacking the computer science background necessary to identify inefficiencies, the tech debt mounts.
I thought of this when the latest jobs report was released, which fell short of expectations but showed a drop in the unemployment rate. The podcast Marketplace described this as a low-hire, low-fire environment, meaning that hiring is anemic, but firing is likewise rare. The one seeming exception was companies claiming to be firing due to efficiencies created by AI (which some have suggested is more posturing than reality). This type of job environment means that a generation of workers will have had less opportunity to build skills and experience through mentoring and doing. Which leads to the question of who will have the skills to resolve the tech debt created by sloppy use of AI and its eventual bugs?
I am not sure, but the brain drain from our own industry has been obvious since the career fallout from the global financial crisis of 2008-2010, only here compounded by robot advisors and fintech solutions (and gamified trading!) I learned early in my career to “beware the crowd at extremes,” which is why we invest in young people in our practice so that we remain competitive for when things inevitably change.