AI and collective memory: does the team still know who knows what?

An AI assistant can become the first point of contact for finding a procedure or preparing a case file. The managerial question is specific: how can this access be maintained without losing sight of the people who understand exceptions and the history behind decisions?

What the literature establishes

Kyle Lewis (2003, Journal of Applied Psychology) developed and validated a measure of transactive memory systems in teams. This concept describes a collective organisation of knowledge: members hold distinct expertise, recognise its credibility and coordinate its use. A team’s memory is therefore more than a collection of documents; it also includes knowing who knows what.

Remembering information or remembering where to find it

Betsy Sparrow, Jenny Liu and Daniel M. Wegner (2011, Science) demonstrated experimentally that expecting future access to information can change what people remember. In their experiments, participants notably remembered where to retrieve information better than the information itself. This mechanism sheds light on delegating memory to an external resource, without demonstrating what happens in a team using generative AI.

The premature conclusion

One might conclude that consulting AI necessarily impoverishes collective knowledge. These studies do not establish that: storing information externally can be useful, and neither article evaluates today’s generative assistants. The managerial hypothesis to examine is more specific: does the tool facilitate access to knowledge while keeping visible the people who can put it in context? (our executive and employee training programmes)

What common usage metrics do not measure

Counting queries or retrieved documents does not establish whether employees can still identify internal expertise. Nor are experiments on individual memory sufficient to predict how working relationships will evolve. Within an organisation, staffing changes, documentation and the allocation of responsibilities can alter this awareness of expertise independently of AI.

A practical check in Lugano

As part of SHR’s programme « Manager à l'ère de l'intelligence artificielle », a team in Lugano can compare its ability to identify the relevant internal contact for equivalent business situations, presented in Italian, before and after a defined period of use. Cases might concern an exception to a banking procedure, a supplier specification in fashion, a contractual term in trading or a customer commitment in a family-owned SME, without entering confidential data into the tool. The proposed measure is the proportion of cases in which an employee names a resource person whose expertise is confirmed by the process owner and explains that choice, distinguishing answers obtained with and without the assistant. This comparison does not prove a causal effect of AI; it checks whether access to answers remains accompanied by access to expertise. To go further: explore the Managing in the Age of Artificial Intelligence training in Lugano, or browse our executive and employee training programmes in Switzerland.

In pictures: Managing in the Age of Artificial Intelligence in Lugano

• managing in the age of artificial intelligence
• Lugano
• research
• training lugano