Catching AI’s mistakes is a mental burden – and a drag on productivity
The time, energy and skills needed to verify AI’s answers to your queries can erase any productivity gains that the technology provides.
The use of AI tools can boost productivity, but workers must spend time and mental effort verifying the accuracy and appropriateness of the output. This verification burden is a double-edged sword, eroding the productivity gains from AI. Cognitive engineers study how people interact with automated decision-support systems, and generative AI verification is a new challenge emerging from the use of this technology.
Research shows that verifying new information is a distinct task from creating it, and it is more cognitively demanding. People tend to overestimate AI's capabilities and believe they are faster when using it, but studies have found no productivity improvements or even slower completion times when verification is included. The burden is greater when AI generates work beyond a person's evaluation capacity.
Verification can feel burdensome in terms of cognitive load, time spent, and motivation. It results in a skill swap, where workers shift from producing to verifying information. To maximize productivity gains, companies must provide workers with the necessary time, training, and authority to verify AI-assisted output.
Written by urgent.news from The Conversation's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.