AI Is Still So Riddled With Errors That It’s Erasing Any Productivity Gains That Would Make It Useful in the Workplace

Wait 5 sec.

Could generative AI be a workplace productivity miracle? Maybe, if only it weren’t for those pesky hallucinations that ruin the whole gambit.At least those are the findings from professor of cognitive engineering at the Georgia Institute of Technology Karen Feigh, who wrote in the Conversation that AI gains in productivity are being offset by the huge amount of time and effort it takes to clean their outputs up.In particular, Feigh points to a phenomenon known as the “verification burden,” a machine learning concept referring to the wasted time humans spend verifying the accuracy of the tech’s work.“At a deeper level, the burden is the cognitive cost of trying to regain authority over work that is partly delegated to a system, while retaining responsibility for the final product,” Feigh explains.The phenomenon probably explains the findings of a recent survey out of Stanford, which found that most workers who’ve been forced to use AI on the job are really just churning out low-effort busywork. Despite giving the appearance of passable reports and memos, this “workslop” inevitably demands more effort from employees down the line.Those workers who do try to check their AI work before they send it on end up wasting more of their own time doing so, a separate analysis found, especially compared to employees who don’t touch AI at all.As Feigh writes, workplace verification tasks are fundamentally different than the act of just creating things ourselves. While drafting a pitch deck generally comes with an obvious end point, double-checking an AI drafted pitch deck is a whole different beast.“You generally know when something you are creating is ‘done,’ or is at least sufficient for the task. When to stop verification is less obvious,” Feigh explains. “A worker must decide what to check and how; they do not know whether an error exists, where it might be or what form it might take.”Put it all together, and it’s no wonder why AI hasn’t revolutionized labor — because it’s simply much more trouble than it’s worth.More on AI and productivity: Remember When Bosses Were Forcing Their Workers to Use AI? They’re Now Realizing They Made a Huge MistakeThe post AI Is Still So Riddled With Errors That It’s Erasing Any Productivity Gains That Would Make It Useful in the Workplace appeared first on Futurism.