Human resources officer recently published an article about the increasing use of different types of quotas and incentives to encourage employees to use AI tools. This is a shockingly bad idea.
There are many reasons. First, if you get people to do something, you’ll get the results you’re measuring, but they’re rarely the fundamental changes you want. Quotas and incentive programs are often driven by self-reports such as “I used 10% more AI this quarter,” but what we really want is the ability to do existing work faster and better, especially the ability to solve new problems.
Let’s assume my leader gave me a quota for this quarter. How would you answer that? Of course you could just make it up, because you know there’s no way to verify your claim. Or, if you feel guilty, you’ll end up using ChatGPT more often, even if you don’t have to. Instead of looking for a caterer for your event, hire AI. It generates things like March Madness bets and draft notes that I probably won’t use. You spend more, but you don’t get any real value.
Let’s say our leadership is a little more sophisticated and as part of my performance review, I’m asking them to document three new ways they’ve used AI this year. I can still make stuff, but it sounds better. We have to believe that what our employees tell us is really an improvement.
Most of the people who claim to use AI these days are just using Copilot or Gemini, which are built into search engines, to do their searches instead. Doing so can save you time, but in my experience it’s not a good idea if you want serious answers based on reliable sources. It is possible to measure faster, but it is very difficult to measure “more accurately”.
As you can see from recent developments, what we’ve been doing so far isn’t working. The current state of AI in businessResearch shows that only about 5% of AI projects actually work and deliver real results for organizations. Deploying AI to improve work outcomes in meaningful ways is a challenging multi-job task. Individual employees can’t do it alone. Quotas and incentives for individuals are no substitute for organizational change.
Opportunities for employees to understand that using AI is effective are more effective than incentives
The big opportunity here is whether employees can use their discretion to find ways to make things better, not just individually but within the workplace group. Consider the huge success of “lean manufacturing”. Employees are now responsible for improving the quality, productivity, and performance of their own work areas, eliminating the need for most supervision in the process. why do they do that? We do it both because we want to make our jobs easier and better, and because we care about the organization and its success. It is a group effort, not an individual effort.
What should you do to increase the effective use of AI in your work? First, stop scaring your employees by talking about how many jobs you want to cut. Top-down and holistic efforts to implement AI won’t work if employees think it’s about killing their jobs.
Second, reduce quotas and incentives for the use of AI. Instead, give some time and resources to the first mover group that has clear ideas about what needs to change. Take successful people and ask others to explain what they did and how they did it. If you don’t have one, look for examples among your vendors. Take them to sit down with the group over lunch or coffee and explain what they did and especially how they did it. If there are no questions, people aren’t listening. To determine what to do, we need to look at examples of change in contexts similar to ours. And the more examples, the better.
Read more: What our queer brethren in the return-to-office debate have to say about the future of work
It’s a good idea to recognize and reward groups that use AI to truly improve their workflows. However, it is unlikely that success will come all at once as planned. We need to plant the seeds of AI by giving imaginative people who are not obviously identified upfront the opportunity and time to try out AI tools and talk to people who have at least some experience with AI. Rather than expecting individual employees to come up with major improvements, it makes more sense to solicit proposals for larger uses that require support from the organization.
Third, this is a place where psychological safety is extremely important. It’s about the need to take some risks and feel unpunished if you fail. If we have permission to try something and believe that if it doesn’t work (perhaps due to lack of cooperation) we will be depressed, we won’t even try.
Finally, even if employees believe that AI will not take their jobs away, if they believe that increased productivity will only result in more of the same work being done, they will have no incentive to use AI. This is currently happening with programs, where AI writes the initial code and humans do the tedious work of checking for mistakes as more and more code is generated. The idea that AI will take over boring jobs and give employees interesting tasks seems like a myth. Yes, we are currently creating a report and an employee is checking it. There is no incentive to use AI unless we understand how it will improve our lives. The idea that you can get them to innovate through quotas and incentives is also a myth.
About the author:
Peter Cappelli is HR Executive’s talent management columnist and a fellow of the National Academy of Human Resources. This article first appeared on HR Executive.
