Have we traded one problem for another? The initial appeal of using AI was to eliminate menial tasks previously done by humans. The "promise" was organizations would save costs and improve efficiencies. But, maybe it's not quite that smooth and easy. 

According to a report recently released from Workday and highlighted on the website CFO.com, inconsistent outputs from large language models are introducing hours of additional work for employees as they sift through data produced by these new AI tools. According to the report, nearly 40% of the time employees have saved using AI tools was lost to "rework", which the companies defined as "correcting, clarifying or rewriting low-quality AI-generated content."

"Instead of reallocating time toward judgment, creativity and decision-making, employees spend it correcting low-quality output. At scale, the pattern compounds, translating into millions of lost hours each year in large organizations", says Workday.

So what to do? The report suggests leaders update job descriptions to "clarify where AI is expected to assist, where human judgment is essential and how success is measured." By partnering with these new tools, instead of expecting them to completely replace a human, organizations are likely to scale efficiencies faster and more successfully.