A latest Harvard Business School research discovered that synthetic intelligence instruments improve employee productiveness and accuracy on sure duties however have a countervailing impact on different similarly-difficult duties exterior a sure “technological frontier.”In a latest working paper known as “Navigating the Jagged Technological Frontier: Field Experimental Evidence of the Effects of AI on Knowledge Worker Productivity and Quality,” HBS and Boston Consulting Group carried out a joint research to research the practicality of AI instruments for practical knowledge-intensive purposes in consulting.The research was carried out by HBS professors Karim R. Lakhani and Edward McFowland III, HBS postdoctoral researcher Fabrizio Dell’Acqua, and researchers from the University of Pennsylvania’s Wharton School, MIT Sloan School of Management, Warwick Business School, and BCG.The researchers discovered that AI capabilities presently cowl an uneven set of abilities they time period a “jagged technological frontier”: exterior of this frontier, AI output will not be correct and even may worsen human efficiency.In the research, 758 BCG consultants got 18 practical consulting duties inside this frontier to trace modifications in employee productiveness and accuracy. The research discovered that, in comparison with staff with out AI entry, those that used GPT-4 accomplished on common 12.2 p.c extra duties, 25.1 p.c faster. Additionally, 40 p.c of the trial group produced greater high quality outcomes.At the identical time, consultants utilizing AI for duties thought-about exterior of the frontier had been 19 p.c much less more likely to produce the proper options in comparison with these with out AI.As a results of the research, researchers distinguished between two completely different patterns of AI integration into these duties: “centaurs” and “cyborgs.”“A centaur is somebody that clearly defines what the human does and what the AI does, and makes use of a human for no matter people are finest at, and the AI for no matter its finest at,” Dell’Acqua mentioned.But most consultants had been noticed to be a “cyborg” — the place consultants have a “fixed interplay with the AI.”“These are each kinds of probably profitable collaborations,” Dell’Acqua mentioned.In one process requiring contributors to investigate retail technique from interview notes and monetary knowledge in a spreadsheet, an analysis of correctness revealed that there was a “fairly massive fall” in efficiency within the group with entry to AI — a discovering that Lakhani attributed to consumer error, not technological shortcomings.“There’s no consumer information and we do not actually know what the frontier is,” Lakhani mentioned, explaining that the frontier is “jagged,” or undefined, which ends up in consumer error.“People use it the mistaken approach. People use it as an data search instrument like Google,” Lakhani mentioned. “This will not be Google.”McFowland highlighted the tradeoff between accuracy and pace when AI is deployed inside organizations with circumstances “exterior the frontier.”“We get productiveness boosts. Things, they bought executed sooner, however getting executed sooner to the mistaken reply in lots of circumstances will not be best — or at the least not preferable — to getting the fitting solutions extra slowly,” he mentioned.The research concluded that understanding the “form and place of the frontier” are essential to optimize the influence of AI on employee productiveness.Though that frontier will transfer ahead as AI progresses, McFowland mentioned we’re nonetheless far off from having the ability to “encapsulate all of human cognition and functionality.”“There’s all the time going to exist issues exterior the frontier with capabilities and infrequently the difficult — or perhaps pernicious — half is that we don’t know what these issues are,” he mentioned.While Lakhani mentioned “it’s going to unlock a ton extra potential for people, for all of us, if you know the way to make use of it,” McFowland remained hesitant about calling AI a “revolution.”“Technology has come alongside and altered how we do issues earlier than, again and again,” he mentioned. “But in hindsight, we acknowledge that they didn’t — they modified our world — however they weren’t as life-altering as we thought they could have been.”—Staff author Tiffani A. Mezitis could be reached at [email protected].—Staff author Camilla J. Martinez could be reached at [email protected]. Follow her on X @camillajinm.
https://www.thecrimson.com/article/2023/10/13/jagged-edge-ai-bcg/
Harvard Business School Partners with BCG on AI Productivity Study | News
