# Voices of the week

**The weekly conversation · 14–20 October 2026**

Ideas worth spending time with.

**The work around the answer**

![Abstract gestural painting of distinct coloured rhythms meeting around an open ivory centre.](/assets/publications/voices/2026-10-20/hero.png)

*Original LecXie artwork · AI-generated*

Five contributions on what AI-supported work still needs from people and institutions: shared context, deliberate practice, credible incentives, reliable completion and rehearsed cooperation.

A good answer can arrive before anyone has worked out what to do with it. These contributions look at the less conspicuous work around AI: how colleagues learn from each other, how people examine a suggestion, what makes improvement feel safe, and whether a proposed action actually happens.

Our connecting question is simple: **what must remain visible, practised and shared if assistance is to become dependable work?** This is LecXie’s reading across different sources, not a joint claim by their authors. A small educational study, a workplace survey, practitioner commentary, a research synthesis and a policy argument offer different kinds of evidence. None supplies a universal recipe.

**In this issue**
- Sarah Wells — What keeps a team thinking together?
- Jason Zagami — Make room for the learner’s judgement.
- Bin Zhang, Yanlin Ma and Xiaolei Qian — Support feels different when success looks zero-sum.
- Nigam Shah and Jonathan Chen — Check that the work happened.
- Fundi Skweyiya-Gushu — Can institutions cooperate under pressure?

## What keeps a team thinking together?

![Forest-green and cobalt paint gestures exchange a coral rhythm across an open ivory interval.](/assets/publications/voices/2026-10-20/wells.png)

*Original LecXie artwork · AI-generated*

**Sarah Wells**  
Independent technology consultant and LeadDev Editor at Large, working on platform engineering, AI governance and technical strategy · UK  
**Collaboration · Commentary · 15 October 2026**

**The argument:** Wells asks what happens to shared context, learning and challenge when engineers increasingly work with agents rather than with colleagues.

**What is new:** Her new editorial agenda puts the social work of engineering alongside questions about AI productivity and technical foundations.

**Why it matters:** Collaboration can develop judgement as well as produce an immediate result.

**LecXie’s practical interpretation:** Before removing a shared review, identify what people learn there and where that exchange will happen instead.

**Evidence & limits:** Practitioner observation and questions for investigation, not a measured decline in team learning. Her consultancy, publishing and speaking interests are visible. The research she discusses predates this week.

[Read the original](https://leaddev.com/leadership/5-shifts-im-watching-as-leaddevs-new-editor-at-large) · [Author biography](https://leaddev.com/community/sarah-wells) · [LinkedIn](https://www.linkedin.com/in/sarahjwells1/)

## Make room for the learner’s judgement.

![Curved blue, ochre and coral paint movements reveal scraped and reworked layers beside an open ivory field.](/assets/publications/voices/2026-10-20/zagami.png)

*Original LecXie artwork · AI-generated*

**Jason Zagami**  
Senior Lecturer, School of Education and Professional Studies, Griffith University; researcher in educational technologies and teacher learning · Australia  
**Learning · Research · 16 October 2026**

**The argument:** Zagami examines a routine in which trainee teachers prompt an AI, evaluate its response and refine their work while documenting their reasoning.

**What is new:** Personalised agents are embedded in ordinary coursework, making the learning process—not merely the finished answer—the object of study.

**Why it matters:** Useful assistance should leave room to practise evaluation.

**LecXie’s practical interpretation:** In a low-stakes training exercise, ask learners to explain one suggestion they accepted and one they challenged. Do not turn private AI conversations into performance surveillance.

**Evidence & limits:** Fourteen participants, one university, twelve weeks; exploratory qualitative evidence, not a causal learning-gain estimate. AI assisted analysis and writing. No relevant conflicts declared.

[Read the original](https://link.springer.com/article/10.1007/s11423-026-10746-4) · [Griffith profile](https://app.griffith.edu.au/phonebook/phone-search.php?format=advanced&name=Jason&surname=Zagami) · [ORCID](https://orcid.org/0000-0003-2003-9546) · [LinkedIn](https://au.linkedin.com/in/jasonzagami)

## Support feels different when success looks zero-sum.

![Aubergine and green paint masses remain in tension while a warm coral and ochre band opens between them.](/assets/publications/voices/2026-10-20/zhang.png)

*Original LecXie artwork · AI-generated*

**Bin Zhang, Yanlin Ma & Xiaolei Qian**  
Bin Zhang and Yanlin Ma: Business School, Sichuan University. Xiaolei Qian: Sichuan International Studies University and HELP University · China/Malaysia affiliations  
**Work design · Research · 14 October 2026**

**The argument:** Perceived AI support was associated with reflective strategies and reported initiative. The support–reflection relationship was weaker among respondents who saw organisational success as one person gaining at another’s expense.

**What is new:** The study examines a boundary condition: the beliefs people bring to an apparently supportive tool.

**Why it matters:** A friendly interface may coexist with reasons to withhold an improvement.

**LecXie’s practical interpretation:** Ask whether sharing an AI-assisted improvement helps the team or threatens someone’s standing; test the answer locally rather than diagnosing individuals.

**Evidence & limits:** Self-reports from 313 employees at six internet firms, collected in three waves. The zero-sum measure uses one item. Associations are not causal evidence; perceived support is not proof of AI goodwill. No funding or commercial conflicts declared.

[Read the original and author record](https://www.frontiersin.org/journals/psychology/articles/10.3389/fpsyg.2026.1931407/full)

LinkedIn profiles for these three authors have not been independently verified.

## Check that the work happened.

![Separated blue paint marks gather into a grounded dark mass with a vivid coral accent.](/assets/publications/voices/2026-10-20/shah-chen.png)

*Original LecXie artwork · AI-generated*

**Nigam Shah & Jonathan Chen**  
Shah is a Stanford Professor of Medicine and Chief Data Scientist at Stanford Health Care. Chen is a Stanford physician-scientist and Associate Professor in Computational Medicine and Biomedical Data Science · United States  
**Evaluation · Research synthesis · 19 October 2026**

**The argument:** Judge an agent by the completed workflow and the changes it makes in the working system, not just the quality of its answer.

**What is new:** Shah and Chen connect clinical and administrative benchmarks around failures between steps: information lost, actions omitted and work left unfinished.

**Why it matters:** An accurate description of an action is not evidence that it occurred.

**LecXie’s practical interpretation:** Define the observable end state before a trial. A service request should have the right owner, authorised action and recorded outcome—not merely a convincing completion message.

**Evidence & limits:** New synthesis of earlier research. Selected tasks, simulated environments and specified models do not establish live hospital outcomes. The headline benchmark scores are not directly comparable. The authors are involved in the work discussed; Shah has company-founding interests, and the synthesis supplies no complete current interests declaration.

[Read the original](https://hai.stanford.edu/news/measuring-what-medical-ai-can-actually-do) · [Shah’s Stanford biography](https://profiles.stanford.edu/nigam-shah) · [Chen’s Stanford biography](https://profiles.stanford.edu/jonc101) · [Shah on LinkedIn](https://www.linkedin.com/in/nigam) · [Chen on LinkedIn](https://www.linkedin.com/in/jonathan-h-chen)

## Can institutions cooperate under pressure?

![Five distinct coloured paint masses exchange traces across an irregular ivory interval.](/assets/publications/voices/2026-10-20/skweyiya-gushu.png)

*Original LecXie artwork · AI-generated*

**Fundi Skweyiya-Gushu**  
Governance and public-policy researcher focused on AI governance, emerging technologies and international security; PhD candidate and former Armscor board member · South Africa  
**Readiness · Policy analysis · 20 October 2026**

**The argument:** Responding to AI-enabled cybercrime requires institutions to practise sharing evidence and making decisions together, alongside formal agreements.

**What is new:** Skweyiya-Gushu proposes recurring African cyber-resilience exercises that expose operational failures, with access rules and independent oversight for shared intelligence.

**Why it matters:** Capable organisations can still fail at the handover.

**LecXie’s practical interpretation:** Rehearse one incident with security, the operational owner and an external provider. Record where information or authority becomes unclear, without using live sensitive data.

**Evidence & limits:** Source-linked policy analysis, not an evaluated intervention. Earlier operations and exercises inform the proposal; they are not new October results. Her former Armscor board role is disclosed.

[Read the original](https://africanarguments.org/2026/10/africas-cybersecurity-gap-is-not-a-treaty-problem-it-is-a-readiness-problem/) · [Author biography](https://africanarguments.org/author/fundi-skweyiya-gushu/) · [LinkedIn listed in her published CV](https://www.linkedin.com/in/fundiswa-skweyiya-662003a)

Her LinkedIn address is sourced from a 2022 published CV; current profile access is unconfirmed.

### Putting the questions together

Imagine a service team using AI to prepare a response to an unusual customer request. The junior colleague explains the reasoning, an experienced colleague challenges an assumption, the team checks the source, and the responsible person verifies that the agreed action reached the correct system. A later review asks whether the person who raised the problem was helped or penalised.

That is an illustrative work-design question, not a reported result. It shows why faster drafting alone cannot tell us whether the whole arrangement is working.

### How we choose

We follow questions as well as names, trace contributions to their original sources, inspect the evidence and disclose relevant interests. The five selections add fresh work published in this coverage period; none repeats a contributor from our first issue. Dates shown are publication dates of the selected pieces, not assumed dates of their fieldwork.

This selection includes an African policy voice and research or practice connected with Australia, China, Malaysia, the United Kingdom and the United States. One contribution does not make Africa—or any other region—adequately represented. Public, accessible and indexed coverage remains uneven. Follower counts, likes and nomination volume have no role in selection.

### Continue the conversation

Previous issue: [7–13 October 2026 · Who controls AI-shaped work?](https://www.lecxie.com/publications/voices-of-the-week/2026-10-13.html)

[Thinkers to watch](https://www.lecxie.com/publications/voices-of-the-week/thinkers.html) remains an unranked, provisional directory. Selection is not an endorsement.

Editorial direction by Dominic Chiappe · research and original artwork with OpenAI.

The illustrations explore each contribution’s ideas; they do not depict the contributors.