Elizabeth Stokoe

@lizstokoe.bsky.social

Professor and Associate Vice President (Impact) at #LSE | Psychologist | Hon FBPsS | Hon Prof @lborouniversity.bsky.social | @IndependentSage.bsky.social Behaviour Group | Conversation Analyst #EMCA | she/her | own views | #AcademicSky

@leonardocarella.bsky.socialReposted by @lizstokoe.bsky.social

In other words, a 10 percent contraction in one of Britain’s most successful exports. If it were steel or fish or other infinitely less profitable sectors, people would be having a meltdown.

@lizstokoe.bsky.social

#EMCA shows that and how humans manage interactions - effectively, less effectively, and sometimes in ways that are mundane yet powerfully poignant - but, like much of interaction, momentary and would not be identified, described, and shared in order end up in a list of "discrete aspects."

8/8

@lizstokoe.bsky.social

"handle inquiries often from recently bereaved customers" - who are being supported by call-takers who are being evaluated moment-to-moment by “The efficiency of AI" which gives "more informed support, coaching and development for colleagues" - based on...? 7/8

@lizstokoe.bsky.social

From the article: "A model from OpenAI has been trained to assess more than 50 discrete aspects of each phone call" - which aspects? It would be interesting to see what crossover there is between these 50 and things that would be endorsed by #EMCA research findings.

6/8

@lizstokoe.bsky.social

From the article: "rates every phone call" with "a % score for interactions between agents and customers"- against what criteria? anything that uses words spoken, length of turns, numbers of turns taken, "silence", "sentiment" etc., is unlikely to be capturing what is happening in the INTERaction.

@lizstokoe.bsky.social

From the article: "automated listening technology" - listening to what? ASR may be accurate about words but social interaction is full of pacing, volume, pitch, intonation, gesture (even when on the phone!). 5/8

@lizstokoe.bsky.social

This includes human-human assessment systems; e.g., mystery shoppers do not make requests etc. in the way real callers do but score service anyway! Given all this, it seems likely that newer technologies are similarly trading off the amassed body of information about what "good" looks like. 4/8

screenshot of article title and abstract Can humans simulate talking like other humans? Comparing simulated clients to real customers in service inquiries Elizabeth Stokoe Rein Ove Sikveland Saul Albert Magnus Hamann Loughborough University, UK William Housley Cardiff University, UK Abstract How authentic are inquiry calls made by simulated clients, or ‘mystery shoppers’, to service organizations, when compared to real callers? We analysed 48 simulated and 63 real inquiry calls to different veterinary practices in the United Kingdom and Ireland. The data were transcribed for conversation analysis, as well as coded for a variety of call categories including reason for the call, call outcome and turn design features. Analysis revealed systematic differences between real and simulated calls in terms of (1) reasons for the call, call outcome and call duration and (2) how callers refer to their pets in service requests and follow-up questions about their animal. Our qualitative analyses were supported with statistical summaries and tests. The findings reveal the limitations of mystery shopper methodology for the assessment of service provision. We also discuss the implications of the findings for the use of simulated encounters and the development of conversational agents
@lizstokoe.bsky.social

One of the reasons I developed CARM was to challenge conversational products and services (e.g., communication training, assessment scripts, guidance, etc.) is because they hardly if ever touch the empirical reality of social interaction and how it works. 3/8

@lizstokoe.bsky.social

In a nutshell, I've yet to encounter a system that is learning/trained via a) actual recorded social interaction; b) transcripts that capture the full range of resources that people use to interact; c) findings from the analysis of interaction. 2/8

screenshot of article Annual Review of Applied Linguistics (2026), 1–16 doi:10.1017/S0267190526100233 What is “conversational” about conversational technologies, products, and services? Insights from conversation analysis Elizabeth Stokoe , Saul Albert and Cathy Pearl Abstract “Conversational” technologies, products, and services are in the headlines more than ever. But what does it mean to be “conversational?” We address this question through the lens of six decades of empirical research in conversation analysis, which has identified and described the foundational structure and interactional machinery of human sociality. We consider not only how tacit notions of “conversationality” manifest in technologies, products, and services, such as role-play, communication training, and chatbots, but also in research methodologies such as focus groups, semi-structured interviews, and laboratory studies – all of which rarely acknowledge how researcher–participant interaction shapes the data collected. Drawing on a range of examples from different institutional settings, we consider how and whether such technologies can or should leverage “conversation” in ways that reproduce “naturalistic” interactions – and ask what might count as “naturalistic” in this context anyway? We argue that if human conversationality becomes a benchmark, then humans themselves will fail tests derived from normative, not empirical, understandings of how social interaction works.
@lizstokoe.bsky.social

There are many stories about AI and 'conversational' products, but this one in particular (about the recording and analysis of probate calls in Co-op Legal Services) had many disappointing, frustrating, and worrying components - from the perspective of conversation analytic research #EMCA.

🧵 1/8

Screenshot of Guardian story - ‘Dystopian’: Co-op becomes latest firm to put staff under AI surveillance Whistleblower says recording and analysis of all probate calls in Co-op Legal Services is oppressive photo: One employee says the system can mean their work is monitored by AI for seven hours a day as they handle inquiries, often from bereaved customers. Photograph: Maskot/Getty Images
@privateeyenews.bsky.socialReposted by @lizstokoe.bsky.social

Who or what is behind the direction of the government’s AI policy? According to minister for artificial intelligence Kanishka Narayan it’s, er, AI!

Older posts