Four students and their tutor at a UK school of nursing had a conversation as part of the series of classroom discussions presented here on this website. In earlier sessions, a significant article was discussed and our understandings and reflections were written up in a collaborative process. On this occasion, wishing to experiment and add to the conversations surrounding AI use, we tried something new: might an online seminar include an AI “member” to everyone’s benefit?
We decided on two roles for the AI. The first and simplest was to take the transcription of our conversation, tidy and thematically summarise it. The first response was very short and did not show which of us had contributed which ideas. The AI was asked to repeat the instruction but “with more detail and preserve the contributors”. This produced something better and it is included at the end of this post. As can be seen, the AI also added some further aspects without being asked: a set of aims for the discussion that it compiled from the first part of our discussion, a series of questions for further discussion and a closing reflection.
Since the full transcript is, like all such documents, full of the normal blunders of speech and at times quite circular, it is very difficult to see a downside to this use of AI. The summary is good and the additions quite useful; the further discussion points are sensible and are evolutions of the discussion we had; the closing reflection seems very fair. The second and third aims we not ones we started with, but which quickly evolved as we began to puzzle over the chosen topic.
The second use of AI was to be as a research partner during our discussion for us to learn about a topic that would be relevant to our module. Prior to this conversation we had practiced a series of principles for university use of AI devised informally by the tutor. These principles were shaped the need to have interactions that are informed and supported by peer-reviewed or other academic sources. The main principles we followed were “boost the sources” and “get out of echo-chambers”. The first of these is simple: insist on knowing the sources that the AI is using, ask it to only use scholarly material, and ensure that conversation ends with published reading material to source through the university library. The second surrounds questions important to critical thinkers long before the use of computers: Who agrees? Who disagrees? Is this text the origin of these ideas? Who else has it influenced, and how?
The tutor deliberately chose a difficult topic, because of a hope that it would stretch all the participants and the test the effectiveness of the AI, too. We set aside two hours for the conversation. The AI was Chat GPT-5, accessed through Microsoft CoPilot. The chosen topic was to seek to understand a translated quotation from Michel Foucault’s Le Pouvoir psychiatrique, from his 1973–74 lectures at the Collège de France, published posthumously in 2003:
«La psychiatrie fonctionne comme pouvoir bien avant de fonctionner comme savoir.»
In English: “Psychiatry functions as power well before it functions as knowledge.”
The group task was to try to understand this quote. To do that we felt we needed to know the following, which we decided upon at the beginning of our session:
Where is this from, where was it published?
What is Foucault’s interest in psychiatry and where does this particular quote sit within Foucault’s ideas?
What was Foucault’s personal background?
Is Foucault the first author to have written on this theme?
Who has been influenced by it and is it still considered important? Are any of these UK psychiatrists or nurses?
Is this theory talking about clinicians? About service-users? Or the role of psychiatry in society?
We opted to have one combined conversation with the tutor sharing his screen rather than all addressing different questions, so that we could all learn together.
The first response was brief and helpful:
AI: This line appears in the foreword or introductory section of the text, where Foucault establishes his central thesis: psychiatry originated not from enlightened scientific insight into mental illness, but from powerful disciplinary mechanisms—asylums, institutional control, surveillance and categorization techniques—that structured how « madness » was socially managed, well before the development of coherent psychiatric knowledge.
Foucault’s lectures argue that psychiatry’s historic role has been shaped predominantly by power structures and disciplinary apparatuses—from Pinel’s reforms to Charcot’s neurology—rather than purely by the advancement of scientific understanding of madness.
Unfamiliarity with Pinel and Charcot led us to ask about Foucault’s use of them, and the AI answer included the term “moral treatment” and “power-knowledge” without explanations, and some more obviously special uses of the word “disciplinary”. Nevertheless, the response did clarify that:
AI: Foucault argues that what we understand as psychiatric ‘truth’—its diagnostic categories, clinical authority, the language of illness—did not emerge simply by observing mental illness. Instead, psychiatric knowledge was constructed within institutional and disciplinary frameworks starting with Pinel in the late 18th century and extending through Charcot in the late 19th century.
We felt this began to partly explain to the original quote – that before psychiatry was knowledge, it was power – literally historically/chronologically before. But we could also begin to see that Foucault was also arguing that it is not just about the sequence of events, but rather that psychiatry was primarily about power, and only secondarily a presentation of knowledge. However, we wanted to know what was meant by “disciplinary frameworks”.
The AI answer, under subheadings of Surveillance, Normalisation, Hierarchical Observation and Examination, and Corrective Techniques, was clear enough, but included the unexplained expression “internalise the gaze”. The tutor felt it best to explain about Foucault’s use of the Panopticon (which was also mentioned, but again, without elaboration) to help teach about “the gaze”, and students seemed to understand this quickly.
Other aspects of the conversation can be seen in the transcript. As we asked for authors who had been influenced by this idea, we were introduced to a number of people, including Pat Bracken and Phil Thomas – who were both given fictitious first names (might “Paddy Bracken” be seen as a little insulting?)! This bizarre moment would have gone unnoticed by the students (although not as they continued to ask to be pointed to actual published work) and shows a very odd feature of AI at present – that is able to be a genuinely useful tool in explaining some fairly complex philosophy with refreshing clarity, but then creates this terrible bit of confabulation. Had a student used this in submitted work they would have looked very foolish indeed.
With the accompanying discussions, our two hours were nearly over but there was time to request five annotated pieces of work for continued study (see below), and these look to be a mixture of key author texts and some by nurses. They are all real references.
The students reflected that the session had been very interesting (one called it “mind-blowing”) and AI had introduced them to a lot of good ideas and authors. They found the topic thought-provoking and saw that it implied that the gap between “care” and “control” is not straightforward. It allowed us to speculate that medical approaches to distress are perhaps dominant because of their political usefulness for all concerned rather than because of science, exactly as Foucault intended us to consider. They were able to move between thinking of the “gaze of the clinic” as being within services and also outwards, into wider society, which felt like the beginning of an important further conversation. Students were all impressed with the thematically organised transcript, which was ready to view just minutes after we had finished the conversation. We did not have time for all of the follow-up questions we could have asked. We should probably have been more consistent in demanding to see the supporting sources of information (this is perfectly possible, but we got rather engrossed in the content).
One of the main strengths of this approach to conducting a seminar using AI as a guide might be that rather than students passively receiving a prepared session, it demonstrates to them some key skills that can help them in their individual learning, too:
- using AI as a tool to quickly gain understanding of a theme
- the research skills of forming deeper, critical questions about a topic
- crafting follow-up questions when presented with new knowledge (essentially, being brave enough to identify what is not yet clear or well-understood, being specific about it and turning it into a question)
- the need to be discerning about sources (the false first names given for two authors at one point really demonstrated the need to be wary and push for accuracy).
- that AI use still needs to lead back to published scholarly material but can be an excellent way of finding it. Writing a few sentences about an idea and asking, “who writes about this?” is a refreshing method of searching for new literature.
The main weakness we encountered seemed to be the level of assumed understanding about Foucault! The AI quickly began to give answers that were peppered with unexplained Foucauldian expressions. In the context of the class this started to feel a bit overwhelming and the tutor had to pause the group interaction with AI and do some more direct teaching. Some would say that is only reassuring – AI has not replaced the teacher yet! One possibility is that the tutor had inadvertently trained his own CoPilot account to write at the level of tutor, not for undergraduates, and that this could have been an instruction given to the AI at the beginning of the conversation (literally, ‘please respond at an undergraduate level’).
One student felt that Foucault had challenged her to see that service-users on wards that she had encountered were often primarily responding to their experience of power, too. This could be seen in distress and frustration, and in conflict within relationships. These, however, were often interpreted by staff as being parts of the character or illness of the person, rather than a response to power.
Another student had clearly been challenged to see that knowledge is power.
As we looked over the transcript, we wondered again if common mental health nursing perspectives, and psychiatric understandings, are less about evidence and really more about the things we need to believe in to cope with the power we have been given. However, we could see that there is a mirror to this: some service users may be attracted to psychiatric understandings for their own needs (which we could be sympathetic about) to control who or what may be deemed responsible for their distress, and for that reason may embrace that power of psychiatry even more readily than mental health professionals. If that is true, it could be said to be another way in which “psychiatry functions as power well before it functions as knowledge”.
The AI generated thematic summary of our conversation, plus reference list, with tutor and student names initialised:
Below is a more detailed thematic summary of the session captured in MSci Extension project.docx, with contributions attributed to the participants.
Session Context
- Date/Time: 7 Nov 2025, ~10:10–11:57 (approx. 1h 47m)
- Participants:
- JG (facilitator; Senior Lecturer)
- OG (student)
- CM (student)
- JHJ (student)
- BS (student)
- Catalyst quote (attributed to Foucault):
“Psychiatry functions as power well before it functions as knowledge.”
- Aims (as co‑constructed in the first 10 minutes):
- Identify where the quote comes from (source, date, context).
- Clarify what Foucault means by power/knowledge, disciplinary frameworks, the gaze, and the panopticon.
- Understand how this thesis relates to Pinel and Charcot.
- Explore whether the idea is original to Foucault and who has been influenced by it.
- Elicit arguments for and against; consider practice implications for mental health nursing.
- Produce an annotated reading list to guide degree‑level follow‑up study.
Thematic Synthesis (with contributors)
1) Power before Knowledge: What the quote implies
- Key idea: Psychiatry’s historical emergence is rooted in institutions and techniques of control (asylums, surveillance, categorisation) rather than in prior, mature scientific understanding.
- Interpretive turn: The phrase “well before” was clarified to mean historically prior (asylums predate the consolidation of psychiatric science) and conceptually prior (psychiatric practice operates as discipline even when framed as knowledge).
- Contributors:
- JG foregrounded the historical and conceptual double-meaning of “well before”.
- OG queried why—if this critique has been around for decades—practice has not shifted more decisively, opening the theme of persistence of power despite critique.
2) Disciplinary Frameworks: What counts as “discipline” in Foucault
- Core components: Surveillance, normalisation, examination, routines/timetables, documentation, and corrective techniques that shape conduct.
- Inside the asylum: Observation, ward routines, risk assessments, and case notes create a field where patients are made legible and manageable.
- Outside the asylum: These logics diffuse to families, schools, law, and culture, shaping social norms of “normal” vs. “abnormal”.
- Contributors:
- CM connected “discipline” to contemporary restrictive practices and observation on wards, asking whether proclaimed “care” sometimes masks control.
- JHJ emphasised the subjectivity of “normal”, challenging uncritical use of the term and highlighting power asymmetries in institutional life.
3) The Gaze & Panopticon: How the clinic “sees” and shapes
- Panopticon (as metaphor): A structure where possible constant observation leads to self‑regulation; people behave as though they are always watched.
- Clinical gaze: Converts persons into cases; gathers signs, measurements, and narratives to fit diagnostic categories and treatment pathways—a technology of power as much as knowledge.
- Everyday examples used in class:
- CM: Speed cameras—behaviour changes when people feel watched; 1:1 observations on wards alter how patients act moment to moment.
- JG: Ward documentation and reports as a surveillance apparatus that shapes behaviour and identity.
- Contribution: All participants helped anchor abstract ideas in concrete, practice‑proximal examples.
4) Pinel & Charcot in Foucault’s arc
- Pinel: “Moral treatment” removed chains yet instituted new regimes of discipline (routines, observation, reform of conduct).
- Charcot: Marks a transition as neurology and, nearby, psychoanalysis reconfigure authority; psychiatry’s domain fragments, but its normalising gaze persists in new locations.
- Contributors:
- JHJ asked how Pinel/Charcot relate to Foucault’s thesis; JG clarified their symbolic roles in the genealogy (from asylum‑centred discipline to clinical specialisations).
5) Power/Knowledge: From genealogy to practice
- Power ≠ a possession; it circulates through relations, techniques, and discourses; knowledge is produced within these relations.
- Implication for nursing: Diagnostic and risk frameworks may produce the very subjects they study (“the psychiatric subject”), demanding reflexivity about documentation, observation, and “evidence‑based” routines.
- Contributors:
- OG: “Knowledge is power” — if we invest in richer knowledge and skills, might we reduce reliance on coercion (e.g., restraint, seclusion)?
- CM: Asked if person‑centred care can be subverted by system‑led priorities; brought trauma, restraint, and ethics into focus.
- JHJ: Pressed the care vs. control tension; how “help” can coincide with mastery.
6) Beyond the hospital: Societal reach of psychiatric norms
- Diffusion: Psychiatry’s classifications and risk logics contribute to societal normalisation, shaping self‑understanding and public policy.
- Contemporary relevance: Medicalisation, diagnostic expansion, surveillance (including digital), and pharmaceuticalisation show how disciplinary power adapts.
- Contributors:
- BS linked the session to assignments and literature; JG noted that continued dominance of a medical model may reflect the utility of its power, not its scientific sufficiency.
Practice‑Facing Insights Raised by the Group
- Care–Control Dialectic (all): Every act of “care” in mental health can have controlling effects; reflexivity is obligatory.
- Observation as Intervention (CM, JG): 1:1 observations, documentation, and risk tools are not neutral; they shape behaviour and identity.
- Restraint & Trauma (CM): Coercive measures can re‑enact harm; challenge necessity and design alternatives.
- Person‑Centred, not System‑Led (JHJ, CM): Keep the person visible even as systems press for efficiency, compliance, and risk minimisation.
- Knowledge to Reduce Coercion (OG): Building de‑escalation, relational, and cultural competencies may shrink reliance on “powerful” fixes (restraint/seclusion).
- Persistent Medical Model (group): Its endurance may indicate institutional usefulness (order, manageability) rather than superior truth‑value.
What Worked / What Needed Support (re: AI use)
- Helpful:
- Gave entry points and a map of concepts (panopticon, gaze, disciplinary power).
- Surfaced follow‑up readings suitable for degree‑level students.
- Limitations:
- Jargon density: Needed human facilitation to make ideas graspable.
- Occasional naming slips and over‑generalities required correction.
- Concepts like biopower/genealogy were introduced but under‑explained.
- Takeaway: Human teaching + AI scaffolding was the productive mix; the group valued examples, drawings/visuals, and practice translations to make Foucault usable.
Readings Identified for Continued Study (as discussed)
- Roberts, M. (2005).The production of the psychiatric subject: power, knowledge and Michel Foucault.Nursing Philosophy, 6(1), 33–42.
- How power/knowledge produces “psychiatric subjects” and what that means for nursing identity and practice.
- Iliopoulos, J.-G. (2022).Foucault understood critical psychiatry.BJPsych Advances, 28(4), 252–261.
- Positions Foucault as offering a deeper critique than anti‑psychiatry; helpful for situating contemporary debates.
- Bracken, P., & Thomas, P. (2010).From Szasz to Foucault: On the role of critical psychiatry.Philosophy, Psychiatry, & Psychology, 17(3), 219–228.
- Compares Szasz and Foucault; clarifies a Foucauldian strand in critical psychiatry.
- Leoni, F. (2013).From Madness to Mental Illness: Psychiatry and Biopolitics in Michel Foucault. In Oxford Handbook of Philosophy and Psychiatry, 85–98.
- Bridges disciplinary power to biopolitics (population‑level management), extending relevance beyond the ward.
- Gastaldo, D., & Holmes, D. (1999).Foucault and nursing: A history of the present.Nursing Inquiry, 6(4), 231–240.
- Early application of Foucault to nursing concepts (clinical gaze, surveillance, resistance), with implications for everyday practice.
Suggested Discussion Questions (for your cohort)
- Where does “care” end and “control” begin? Analyse a recent clinical scenario using the care–control lens.
- Observation as power: How do documentation and risk tools shape what a person can say or become on a ward?
- Alternatives to coercion: Which skills (relational, cultural, de‑escalation, environmental) can realistically displace restraint/seclusion?
- Person‑centred or system‑led? Identify system pressures that override person‑centredness and propose team‑level mitigations.
- Beyond the hospital: Where do you see psychiatric norms at work in schools, social media, employment, or criminal justice?
Closing Reflection
The group moved from a single arresting sentence to a multi‑layered understanding of Foucault’s thesis: psychiatry’s disciplinary power (surveillance, normalisation, examination) historically and presently prefigures and undergirds its knowledge claims. Participants repeatedly returned to practice ethics—restraint, observation, documentation, and person‑centredness—showing that Foucault is not merely a theorist of the past but a lens for current dilemmas in mental health nursing.
Critical Mental Health Nurses' Network 
Leave a comment