Imagine you run into Warren Buffett. He gives you a stock tip. You listen; he’s the Oracle.
Now imagine the older gentleman bagging groceries gives you a stock tip. You probably don’t listen.
But would you listen if you learned he works part-time simply because he likes interacting with people? And that he spent his entire pre-retirement-career running a hedge fund, and his investment accuracy rate remains staggeringly high?
In that case, you probably would.
While the quality of advice, input, and information is all that should count, it’s natural to also consider the source. Authority, experience, intelligence… those things matter.
And the source is artificial intelligence.
In a 2021 study published in Strategic Management Journal, researchers used AI to track employee behaviors and recommend performance improvements. The resulting feedback was more accurate, more individualized, and more relevant for individual employees. Employee performance improved.
Until the employees were told the feedback they received had been generated by AI.
Then their performance actually dropped below pre-study levels — even though the “computer” feedback provided “higher quality feedback in terms of greater breadth and depth than do human managers, which in turn increases employees’ learning and performance.”
And even though the job performance of employees who received “computer” feedback improved by 12 more than those who received “human” feedback.
As the researchers write:
We find that employees to whom AI feedback is disclosed tend to have lower trust in the quality of the feedback and higher concerns over job displacement risk, both of which impede their learning and job performance.
… we demonstrate a negative “disclosure effect” (among) the employees who are informed of receiving AI feedback.
Same information. Different source. Turns out the source matters.
Especially if the source is AI.
Interestingly, the “disclosure effect” made less of an impact on more experienced employees. Partly that might be due to the fact people assume AI will someday eliminate their jobs, a fear the researchers call “perceived displacement risk.” (The longer I’ve been with a company, the more “protected” I probably feel.)
Granted, you’re probably not using AI to track employee performance and make specific improvement recommendations. (At least not yet.)
But you should consider how you deliver performance feedback, especially where the source of that feedback is concerned.
For example, research shows a highly competent boss — one who excels at “ability to get the job done” and “employee development” — has by far the largest positive influence on employee job satisfaction.
As the researchers write, “If your boss could do your job, you’re more likely to be happy at work.” And more happy about — and willing to listen to — the feedback you receive.
But that doesn’t always apply to feedback provided by highly skilled “outsiders.” When I worked in manufacturing, a machine operator from another plant would occasionally be brought in to train us. In terms of job skills, they were more proficient, more experienced, more everything than we were.
But we didn’t listen. While the information was great, we were too hung up on the source.
Sure, they might have been great back wherever they came from… but what they did know about working in our plant? (What does a computer know about doing my job?)
Feedback should be source-independent; the quality and relevance of the information I receive should be all that matters.
But in reality, the source also always matters. Whether the source is AI. Or processes borrowed from other companies. Or targets set through benchmarking. Or feedback provided by a new supervisor or manager.
Or feedback given by you… instead of the person whose opinion is most likely to be trusted and respected by the person who receives that feedback.
Because no matter what the source, feedback isn’t feedback unless the input or advice you provide leads to actual improvement.
Otherwise… it’s just words.