Dear Springboard: We’ve been hearing a lot about generative AI. Who’s getting the most out of it and what’s the overall impact? Sign me, AI Curious Dear AI Curious: The adoption of AI was like an explosion – a huge burst of energy but ultimately not that focused. Consider that ChatGPT was launched November 30, 2022, and it reached 100 million users in 2 months! Compare that to Facebook which took three years to reach 10 million after its launch in 2004 and the iPhone two years to reach 10 million from its 2007 launch. Given this reception and the promise of such powerful technology, the investment in AI – the technology and the data centers—has itself been explosive. A recent article in the Wall Street Journal reported that the scale of investment—hundreds of billions of dollars—is unprecedented. Many of the people interviewed for the article noted it’s hard to see how this outlay will be profitable, at least in the near or even medium term. The rush to investment has been attributed to greed, ego and, ultimately, FOMO.  All the big players want to be a part of the Next Big Thing. While there has been a high rate of adoption of AI at businesses, a recent MIT Media Lab study found that 95% of organizations have not seen a measurable return on their investment. Part of this could be that usage and business models need to catch up with what is available. Another reason could be how generative AI is most used. Many employees are using AI to do their work for them. Based on research conducted in the first half of the year, employees are creating “passable looking work that ends up creating more work for their coworkers,” wrote Kate Niederhoffer (et al) in a recent post on Harvard Business Review. They call this subpar work product “workslop” and they define it as “AI generated work content that masquerades as good work, but lacks the substance to meaningfully advance a given task.” Workslop is incomplete, unhelpful or otherwise inadequate so that the recipient has to “ . . . interpret, correct, or redo the work.” The authors conducted their survey of US-based employees across industries that found that 40% reported having received workslop in the last month.  What’s more, those who have received workslop say that an average of 15% of what they receive is workslop.

Image AI Generated

What we’re seeing is a lazy way to increase one’s productivity, but at a cost to coworkers and the organization.

In the survey, recipients of workslop reported spending nearly two hours with each instance of workslop.

Workslop also has negative social and emotional impact. Recipients need to diplomatically respond, when “53% report being annoyed, 38% confused and 22% offended.”

Not surprisingly, about 50% of the people who received this subpar work saw their “colleagues who sent workslop as less creative, capable and reliable than they did before receiving the output.” Plus, 42% saw them as less trustworthy and 37% as less intelligent. And, the negative impressions were even tougher on women.

Over time, workslop will undermine collaboration and erode trust.

Plus, I personally worry about an atrophy of critical thinking skills.

We’re seeing the same approach to taking short cuts at work as that among college students who opt to have AI completely do their assignments.

Given human nature, these shortcuts are not surprising.  When I’m doing my daily word puzzles online, I’ll confess I sometimes ask for a hint rather than make the effort to reason it out, all the while rationalizing my choice.

Who’s taking the shortcuts and who’s toughing it out?

Research has identified two types of users: pilots vs passengers.  Pilots have a combination of high agency and high optimism and they “use gen AI 75% more often than passengers and 95% more often outside of work.”

It’s not just their heavy usage but that they use it to enhance their own creativity. They use AI as a tool to complement their own efforts.

Passengers, on the other hand, are more likely to use AI to avoid doing the work.

This begs the question of how leaders can manage and direct the integration of AI.

What we’re seeing is an increase in white collar efficiency for some individuals (with a cost  to some of their downstream colleagues) but not the ROI that is hoped for and necessary.

The solution needs leadership, management, and disciplined execution.

Next time, we will look at a leadership-based model proposed by consultants at ghSMART, a highly respected management consulting firm.

For a meta view of what the big players are trying to build, here is another link.