AI improves performance for both individuals and teams, new research finds. Do businesses draw the right lessons from that?
Hello, welcome to the eyes of AI. This edition suggests that AI could become a team player… Openai promotes COO, CEO Sam Altman shifts focus… Apple shakes AI teams.
Evidence of a positive impact on AI productivity continues to grow. However, while many executives ultimately see AI as an alternative to human labor, they hope to ultimately automate tasks and save staff, data suggests that this is not the best way to think about technology. Yes, in some cases, AI can fully automate some tasks. However, in most cases, today’s AI systems include so-called “AI agents.” Salesforce, ServiceNow, Microsoftand Google – not capable or reliable enough to do this. Instead, AI systems should be considered as complementary to human labor. It’s a way to improve people’s performance and not replace them.
The latest support for this view comes from Attractive research It is published as a working paper for the research repository SSRN by Wharton Business School at the University of Pennsylvania, ESSEC Business School in France, Consumer Product Title Procter & Gamble, and Harvard, Harton University at Harton University. (The author includes Ethan Mollick from Wharton, who attracted huge social media for tips on how to effectively use AI in Business.)
In 2024, researchers held a one-day virtual product development workshop at P&G. Famous use-This time we’ll add an AI twist. In particular, the workshop included the “seed” stage of product development. This is to brainstorm many possible new product ideas and incubate them to a point where you can decide whether to test them at a larger scale. P&G usually assigns two-person teams consisting of one commercial operational person and one R&D expert to cooperate with brainstorming ideas. In this case, the researchers took 776 P&G employees from commercial and R&D and randomly assigned them to do one of the following: It works on its own, but with access to the generator AI assistant based on Openai’s GPT-4 model. Works for a normal two-person brainstorming team consisting of one commercial and one R&D person. Or you’ll work in a normal two-person configuration, but you’ll have access to the AI assistant.
The group was then tasked with coming up with new ideas for consumer products in the various P&G departments where they worked (baby care, women’s care, grooming, oral care). These ideas were evaluated by human judges, both in their relevant business and technology expertise.
AI improves individual performance considerably
Two heads are generally better than one, so it’s not surprising that an individual who works alone and has no access to AI has had the worst. However, individuals with AI support have been found to be superior to two-person teams with no AI, on average. In fact, these AI-assisted individuals’ performance was not statistically superior to the two-person teams using AI. This allows us to conclude that AI could indeed be a substitute for human labor. Companies like Procter & Gamble allow two-person product teams to be reduced to a single individual brainstorming with the help of AI.
Individuals using AI also had other major benefits. Individuals who worked with AI were able to work faster. Compared to people who work without AI, they spent over 16% less time coming up with ideas, but teams working with AI were about 12% faster.
Handling AI was better than “bowling alone.” Individuals reported more positive and negative emotions than unsupported lonely wolves during the product idea process.
Importantly, people who work alone tended to come up with ideas that suited mostly professional silos. This is commercial people who support product innovations primarily with regard to new commercial ideas (changes in brands, packaging, or marketing strategies). However, when supported by AI, these individuals achieved a mixed approach by combining both technological and commercial innovations, as did human pairing. “This suggests that AI will provide services not only as information providers but as an effective boundary mechanism, which will help experts tackle the issues more comprehensively across traditional domain boundaries,” the researchers write.
Help your team be extraordinary
However, before jumping to the conclusion that AI should be used to reduce team size, it is important to point out perhaps the most interesting findings of the entire study. The two teams working with AI have generated far more ideas that human experts have been rated as “exceptional.” Additionally, AI-supported human teams reported the most enjoyment of working on tasks compared to other groups.
A blog about the survey resultsMollick writes that “organisations primarily view AI as just another productivity tool, such as better calculators and spreadsheets,” while employees “used AI as well as everyday productivity tasks for critical thinking and complex problem solving.” AI could be considered another member of the team (as a collaborator). “Companies that focus solely on improving efficiency from AI will not only want to share their discoveries with fear of making themselves redundant, but will also miss the opportunity to think bigger about the future of work,” he wrote. He not only tried to automate existing processes, but also encouraged organizations to rethink their work and management structure.
I’m sure this is correct. Unfortunately, the temptation for many managers is to grab the obvious labor and time savings that AI offers, as there is a clear and immediate return on labor savings. Brave executives need to insist on maintaining people properly, but using AI to allow them to be exceptional.
So, there’s the rest of this week’s AI news.
Jeremy Kahn
jeremy.kahn@fortune.com
@jeremyakahn
Before you reach the news, if you want to know more about how AI affects your business, the economy, and our society (and probably so, given you’re reading this newsletter), consider joining the Fortune Brainstorm AI London 2025 conference. The meeting will be held from May 6th to 7th at the Rosewood Hotel in London. Verified speakers include Cohere CEO Aidan Gomez and MasterCard Chief Product Officer Jorn Lambert. eBay Nitzan Mekel, Chief Executive Nitzan Mekel, Shaun Maguire, partner at Sequoia, looked at others such as high-tech analyst Benedict Evans. Of course I’m there. I’ll see you there too. You can apply to attend here.
And if I miss you in London, why not join Singapore on July 22nd and 23rd to join Fortune Brainstorming AI Singapore? Learn more about the event here.
This story was originally introduced Fortune.com