Truace
Subscribe

Are you ready for a cybernetic colleague? How to use AI in the workplace without dumbing down

Artificial intelligence is set to transform our world, but how those changes will affect the workplace is still unclear. As generative AI (gen AI) – such as ChatGPT and Claude – is implemented across organisations and businesses worldwide, managers are experimenting to see how the technology can increase efficiency, improve innovation and boost profits. “AI is at a fascinating, fuzzy and experimental stage where managers are trying everything. They are throwing darts to see what hits,” says Hila Lifshitz, profes…

By the Truvace desk2 min read1 sourceJournalismTRV-2026-1405· cite & verify
Are you ready for a cybernetic colleague? How to use AI in the workplace without dumbing down

In brief

Researchers at Warwick Business School studied 776 professionals at Procter & Gamble and found that generative AI can work as a "cybernetic teammate", contributing creatively to projects. Teams with an AI member were far more likely to produce top-ranked ideas, and did better at creating new products than individuals using AI or teams without it.

The same academics warn that heavy reliance on gen AI can narrow thinking, since everyone draws on similar information, and can erode the critical thinking organisations build. How far these risks play out depends on how firms train staff and structure AI use, which the article presents as suggestions rather than tested results.

Main points

  1. The P&G study covered 776 professionals and found AI acted as a proxy team member that made creative contributions.
  2. Teams with AI outperformed both individuals with AI and teams without AI at creating new products, according to Lifshitz.
  3. Shared access to similar information can create an "ideation bubble" that reproduces historical, standardised knowledge.
  4. Chater proposes adding friction, such as requiring users to upload their own attempt before the AI answers.
  5. At Lloyds Banking Group, augmentation keeps human judgment in high-stakes customer cases such as fraud.

The gain

In a study of 776 Procter & Gamble professionals, human teams with an AI "team member" were three times more likely to produce ideas ranked in the top 10% by peers than teams without AI.

The rundown

The positive evidence comes from a study involving Procter & Gamble staff, where the AI was more than a research resource and behaved like a team member. Peer ranking of ideas was the measure, and teams with AI were three times more likely to reach the top 10%.

The risks center on homogeneity and cognitive offloading. Because AI learns new ideas and makes them accessible to all, Chater describes a possible "tragedy of the commons" in which creators lose attribution and nobody is incentivised to use the technology properly.

Suggested remedies include demanding diverse viewpoints and cross-sector comparisons, adding friction to prompts, and using augmentation. Karunesh describes contact centre staff using gen AI to summarise a customer's contact history while humans handle sensitive cases such as fraud.

Sources

  1. JournalismThe Guardian2026-10-09

The debate