Hey and welcome to Eye on AI. On this version: A brand new research suggests AI could be a group participant…OpenAI promotes its COO whereas CEO Sam Altman shifts focus…Apple shakes up its AI group amid frustration over delayed Apple Intelligence options…a revolutionary new AI climate forecasting technique…and AI transforms structure.
Proof of AI’s optimistic impression on productiveness continues to mount. However whereas many executives view AI as in the end an alternative choice to human labor, hoping it should ultimately totally automate duties and save on headcount, the info means that this isn’t one of the best ways to consider the expertise. Sure, in a couple of circumstances, AI can totally automate some duties. However typically, immediately’s AI methods—together with the so-called “AI agents” from the likes of Salesforce, ServiceNow, Microsoft, and Google—aren’t but succesful or dependable sufficient to do that. As a substitute, AI methods must be considered a complement to human labor—a technique to raise the efficiency of individuals, to not substitute them.
The newest help for this view comes from an enchanting research by a gaggle of researchers—from Harvard, the College of Pennsylvania’s Wharton College of Enterprise, ESSEC Enterprise College in France, and client merchandise large Procter & Gamble—and printed as a working paper on the analysis repository SSRN. (The authors embrace Wharton’s Ethan Mollick, who has attracted an enormous social media following for his recommendations on the right way to use AI successfully in enterprise.)
In 2024, the researchers carried out a one-day digital product improvement workshop at P&G, with the method designed to reflect the one which client merchandise behemoth famously makes use of—besides this time with an AI twist. Particularly, this workshop concerned the “seed” stage of product improvement—which is about brainstorming a number of attainable new product concepts and incubating them to the purpose the place a call will be made on whether or not to check them at a bigger scale. P&G usually assigns two-person groups consisting of 1 Industrial operations individual and one R&D knowledgeable to work collectively on brainstorming concepts. On this case, the researchers took 776 P&G workers from Industrial and R&D and randomly assigned them to do one of many following: work alone; work alone however with entry to a generative AI assistant based mostly on OpenAI’s GPT-4 mannequin; work within the common two-person brainstorming group consisting of 1 Industrial and one R&D individual; or work within the common two-person configuration however with entry to the AI assistant.
The teams have been then tasked with arising with new concepts for client merchandise within the varied P&G divisions wherein they labored (child care, female care, grooming, and oral care). These concepts have been then assessed by human judges with each related enterprise and expertise experience.
AI lifts particular person efficiency—by rather a lot
Two heads are usually higher than one, so it’s maybe not stunning that people working alone and with out entry to AI did the worst. But it surely turned out that people assisted by AI carried out, on common, higher than two-person groups with out AI. In reality, the efficiency of those AI-assisted people was not statistically higher than two-person groups working with AI. This may lead one to conclude that AI can certainly be a very good substitute for human labor—enabling an organization like Procter & Gamble to scale back its two-person product groups to only single people brainstorming with the assistance of AI.
There have been another large advantages to the people working with AI, too. People working with AI have been in a position to work quicker—taking greater than 16% much less time to provide you with an concept in comparison with folks working with out AI, whereas groups working with AI have been about 12% quicker.
Working with AI was additionally higher than “bowling alone”—people reported extra optimistic feelings and fewer unfavorable ones in the course of the product ideation course of than the unassisted lone wolves.
Importantly, folks working alone tended to provide you with concepts that match primarily into their skilled silos—business folks favoring product improvements that have been principally about novel business concepts (adjustments in branding, packaging, or advertising and marketing technique) whereas the R&D specialists favored technological improvements. However when assisted by AI, these people achieved blended approaches, combining each technical innovation and business innovation—identical to the human-human pairings did. “This suggests AI serves not just as an information provider but as an effective boundary-spanning mechanism, helping professionals reason across traditional domain boundaries and approach problems more holistically,” the researchers wrote.
Serving to groups to be extraordinary
However, earlier than you leap to the conclusion that AI must be used to scale back group sizes, it is very important level out maybe essentially the most attention-grabbing discovering of the entire research: The 2 individual groups working with AI produced much more concepts that the human specialists rated as “exceptional”—the ten% that they judged more than likely to result in really breakout merchandise. And the human groups assisted by AI additionally reported essentially the most enjoyment from engaged on the duty, in comparison with the opposite teams.
Running a blog concerning the findings, Mollick wrote that “organizations have primarily viewed AI as just another productivity tool, like a better calculator or spreadsheet,” however that workers have been typically utilizing “AI for critical thinking and complex problem solving, not just routine productivity tasks.” AI might be seen as one other member of the group—as a collaborator—not simply one other software, he wrote. “Companies that focus solely on efficiency gains from AI will not only find workers unwilling to share their AI discoveries for fear of making themselves redundant but will also miss the opportunity to think bigger about the future of work,” he wrote. He inspired organizations to reimagine work and administration constructions, not simply search to automate present processes.
I’m positive that is right. Sadly, the temptation for a lot of managers can be to seize on the labor and time financial savings AI presents, since there may be an apparent and fast pay-off in labor financial savings. It would take braver executives to argue for protecting folks in place however utilizing AI to empower them to be distinctive.
With that, right here’s the remainder of this week’s AI information.
Jeremy Kahn
jeremy.kahn@fortune.com
@jeremyakahn
Earlier than we get to the information, if you happen to’re inquisitive about studying extra about how AI will impression your corporation, the economic system, and our societies (and given that you just’re studying this article, you most likely are), please think about becoming a member of me on the Fortune Brainstorm AI London 2025 convention. The convention is being held Might 6-7 on the Rosewood Resort in London. Confirmed audio system embrace Cohere CEO Aidan Gomez, Mastercard chief product officer Jorn Lambert, eBay chief AI officer Nitzan Mekel, Sequoia companion Shaun Maguire, famous tech analyst Benedict Evans, and lots of extra. I’ll be there, after all. I hope to see you there too. You may apply to attend right here.
And if I miss you in London, why not think about becoming a member of me in Singapore on July 22 and 23 for Fortune Brainstorm AI Singapore. You may be taught extra about that occasion right here.
This story was initially featured on Fortune.com
