How can organisations integrate AI into enterprise systems without losing sight of human needs?
In a new publication, Beni Suranto, Åsa Cajander and Ramesh Lal explore tensions that can arise when AI becomes part of collaborative UX design. Based on a structured review of previous research, they identify six recurring tensions, including fairness versus bias, automation versus human agency, explainability versus usability, and inclusivity versus standardisation.
The authors propose the HCAI CUX framework, which connects these tensions to five principles of Human-Centered AI: transparency, user empowerment, fairness and inclusion, iterative co-creation, and accountability.
The framework provides practical guidance for designing AI-supported enterprise systems that are not only efficient but also transparent, inclusive, and accountable.
The publication is written by Beni Suranto, Åsa Cajander and Ramesh Lal and published in Procedia Computer Science as part of the Eighth Information Systems International Conference.
A new study published in DIGITAL HEALTH explores which information patients consider especially sensitive in their electronic health records.
The study is based on survey responses from 4,459 users of the Finnish national patient portal My Kanta. Mental health information was by far the most commonly mentioned type of sensitive information. Patients who had received mental health care were also considerably more likely to report that some health information was especially sensitive compared with patients who had received other types of care.
An important finding was that sensitivity was not only about the type of information. Patients also described how sensitivity depends on context, including who can access the information, how it might be used, and whether it could have consequences for future care, employment, insurance, or how they are treated by others.
The findings highlight the importance of careful documentation and thoughtful design of patient portals. The authors suggest that patients could be given greater possibilities to limit access to certain information or mark information as particularly sensitive. They also emphasize the importance of involving patients and health care professionals when designing patient accessible health records.
The study was conducted by Saija Simola, Sari Kujala, Anna Kharko, Josefin Hagström, Charlotte Blease, Åsa Cajander, Rose Mharie Åhlfeldt, Bo Wang, Bridget Kane, and Maria Hägglund.
Digital systems at work are often procured, designed, and evaluated one system at a time. But employees rarely work that way. Their daily tasks typically involve moving between several systems, tools, and workflows, which together shape both performance and well-being.
In a new article in Behavior & Information Technology, Marta Larusdottir and Åsa Cajander examine three design methods that can help broaden this perspective: Contextual Personas, Contextual Think Aloud, and Vision Seminars. Their analysis of six empirical studies shows that the methods can help designers understand not only how people use individual systems, but also the wider organizational, social, and cognitive conditions of their digital work environment.
The three methods can support different parts of development. Contextual Personas can help understand users and their work early in the process. Contextual Think-Aloud can be used to evaluate how a system fits into real-world work. Vision Seminars can support discussions about future work practices and the wider digital environment.
The findings also point to challenges, including the risk of stereotypes in personas, difficulties for some users in thinking aloud, and the time and resources required for Vision Seminars. Still, the study suggests that moving beyond a single system-thinking approach is both possible and important when designing digital workplaces that support people in their actual work.
Authors: Marta Larusdottir, Reykjavik University, and Åsa Cajander, Uppsala University.
The course serves as a broad introduction to the research field of the graduate school. It provides doctoral students with common ground for understanding work environment and wellbeing in the context of a digitalised school sector.
Digitalisation is here understood in a broad sense, including traditional digitalisation, future automation, and the use of AI systems. The course focuses on definitions of the work environment, basic requirements, roles and responsibilities, laws and regulations, and their effects on work efficiency, health, wellbeing, safety, sustainability, and preventive work. Active student participation is required through seminars and assignments.
This is a basic course, aimed at doctoral students in subjects relevant to the LärA-AI graduate school.
After completing the course, the doctoral student should be able to:
Describe central concepts, regulations, actors, and responsibilities related to work environment and occupational health.
Discuss physical, social and organisational, cognitive, and digital dimensions of work environment.
Analyse how digitalisation, automation, and AI may affect work organisation, wellbeing, professional autonomy, trust, ethics, and sustainability in the school sector.
Relate work environment issues to relevant research literature.
Formulate research questions concerning work environment and wellbeing in a digitalised or AI-transformed school sector.
Course contents
Seminars, including the following topics:
What is the work environment and occupational health?
Definitions, laws, regulations, actors, responsibilities, etc.
What is a digital work environment?
Physical work environment
Social and organizational work environment
Cognitive work environment
Stress, health and wellbeing
Effects of digitalization and automation
Usability and user experiences
Equal opportunities and gender
Engagement, motivation, work satisfaction, autonomy, trust, ethics
What is new/different when it comes to AI vs. “traditional digitalization”?
Development models for healthy, sustainable work
Methods for the evaluation of the work environment
Systematic Work Environment Management (SAM, SWEM)
Assignment:
The examination consists of active seminar participation and a written orienting essay. In the essay, the doctoral students select a work environment issue relevant to the digitalized or AI-transformed school sector, motivate its relevance, relate it to selected course literature, and formulate possible research questions for further study.
Course schedule
The course will take place from September to November 2026. Some seminars will take place on campus, while others will be offered in hybrid or online format. The course start will be on campus, September 15. Exact dates for all seminars will be announced later.
Literature
Relevant literature, including books, scientific articles and reports will be specified as reading material for each seminar.
Course responsible: Bengt Sandblad and Viktoria Rubin Department of Information Technology, Uppsala University bengt.sandblad@it.uu.se and viktoria.rubin@it.uu.se
To register, please email Bengt Sandblad no later than 25 August 2026. The course has a maximum of 20 participants. Doctoral students from other universities are warmly welcome to apply.
I ett nytt avsnitt av UppTalk deltar Åsa Cajander från HTO och en av föreståndarna för forskarskolan LärA AI. Samtalet handlar om hur AI påverkar skolan, läraryrket och lärares arbetsmiljö.
I intervjun diskuterar Åsa Cajander och Mats Daniels hur generativ AI kan förändra både undervisningens innehåll och organisering. AI kan skapa nya möjligheter för lärare, till exempel genom stöd i planering, återkoppling och administration. Samtidigt väcker tekniken viktiga frågor om ansvar, kompetens, etik och arbetsbelastning.
Samtalet knyter an till forskarskolan LärA AI, Läraryrkets arbetsmiljö i en AI transformerad skola. Forskarskolan undersöker hur AI förändrar lärares arbete och vilka nya kunskaper som behövs för att förstå och utveckla en hållbar digital arbetsmiljö i skolan.
AI och automatiserade triagesystem blir allt vanligare i svensk primärvård. Förhoppningen är ofta att tekniken ska effektivisera arbetet, förbättra tillgängligheten och avlasta vårdpersonalen. Men hur påverkas egentligen sjuksköterskornas arbete när AI blir en del av vardagen?
I en ny studie har vi intervjuat 29 sjuksköterskor som arbetar med ett AI-stött triagesystem där patienter beskriver sina symtom digitalt, får automatiskt genererade frågor och därefter kommunicerar med sjuksköterskor via chatt.
Resultaten visar att AI inte ersätter sjuksköterskornas kompetens. Snarare förändras arbetet.
Många sjuksköterskor beskriver hur de får ett nytt, ofta osynligt, arbete med att tolka, kontrollera och ibland korrigera systemets bedömningar. Trots att AI-systemet sorterar och sammanfattar information krävs fortfarande professionell bedömning för att avgöra vad som är relevant, vad som saknas och om systemets prioritering verkligen är rimlig.
Studien visar också att nya digitala arbetsuppgifter tillkommer. Sjuksköterskorna behöver navigera mellan flera olika IT-system, hantera stora mängder textbaserad information och ofta hålla flera patientärenden igång samtidigt. Det ställer höga krav på både digital kompetens och koncentrationsförmåga.
Samtidigt beskriver många att de förlorar något viktigt när patientkontakten sker via text istället för via telefon eller fysiska möten. Tonfall, tvekan, oro och andra nyanser som kan vara avgörande i en bedömning blir svårare att fånga. Flera sjuksköterskor berättar att de behöver arbeta aktivt med sitt språk för att kunna förmedla empati, lugn och trygghet i skrift.
En annan viktig fråga handlar om arbetsmiljön. Flera deltagare upplevde att de fick mindre kontroll över sitt arbete när systemet styrde arbetsflödet. Chattarna fortsätter att komma in även under perioder med hög belastning, och möjligheterna att själv styra arbetstakten eller ta pauser upplevdes som mindre än i traditionell telefonrådgivning. Att hantera flera parallella patientdialoger samtidigt beskrevs som mentalt krävande och stundtals stressande.
Studien visar också att införandet av AI påverkar relationer på arbetsplatsen. När mer arbete sker digitalt och på distans riskerar spontana möten och informellt kollegialt stöd att minska. Samtidigt lyfter sjuksköterskorna fram att digitala möjligheter att rådfråga kollegor och läkare kan skapa nya former av samarbete.
Det kanske viktigaste budskapet från studien är att AI i vården inte bara är en teknisk fråga. När nya system införs förändras också arbetsmiljön, yrkesrollen och förutsättningarna för att ge god vård. För att digitaliseringen ska bli hållbar behöver vi därför inte bara fråga vad tekniken kan göra, utan också vilket arbete den skapar för dem som använder den varje dag.
In a new paper, Andreas Bergqvist, Niklas Humble, Mats Daniels, and Åsa Cajander investigates how AI influences work engagement among IT professionals in the Swedish IT-sector. 28 qualitative interviews were studied through the job demands resource model. The findings show how emerging technologies such AI can contextually affect both resources and demands in the work practices and workplaces at the same time. On one end it can reduce repetitive tasks, support learning, and highlight meaningfulness of work, while, on the other, it can lead to cognitive overload and concerns about reliability and skill relevance. These findings suggest that the job demands resource model is insufficient to fully account for the dynamic nature of how digital technologies co-constitute work practices contextually through mediated interactions as part of a socio-technical system. This highlights the need to adapt the conceptualization of work engagement to handle the nuanced ways that both direct and indirect interactions with technologies shapes our practice and work. Framed through the need for theories on UX@work, we call this Digital Work Engagement, a positive and fulfilling user experience of vigour, dedication, and absorption based on the worker’s interactions with and relation to technology in the workplace. Based on this, the paper includes implications for how both managers and designers can continue to foster work engagement through how they develop and integrate emerging technologies with respect for professional growth, autonomy, and support.
We gratefully acknowledge that the study and research project was supported by funding from Afa Försäkringar, Sweden [grant number 220244].
As a parallel, the image at the start of this post is AI-generated, or at least partially. I thought I could get it generated ready to post directly. But after sitting an hour prompting, I had a chat of failed attempts. The face changed between the image, the robot got googly eyes or stopped holding anything, and computer screens got flipped the wrong way. In the end, I took my favorite parts of different attempts and cut and pasted them together to fit what I intended. It became self-induced workslop. Next time I will probably just browse for a creative commons photo that sort of fits instead. So in a way, similarly to what you can read in the paper, the AI was simultaneously useful to me and induced more work and frustration.
Informal caregivers (ICs) play a vital role in supporting individuals with head and neck cancer. However, they are often expected to manage complex medical, emotional, and practical responsibilities with limited preparation. This gap between expectations and available support highlights a clear need for interventions that go beyond providing information and instead actively strengthen caregivers’ well-being and sense of preparedness.
In our recent study, co-authored by Awais Ahmad, Åsa Cajander, Ulrica Langegård, Birgitta Johansson, Anna Henriksson, Mona Pettersson, Waqas Ahmad, and Ylva Tiblom Ehrsson, and published in JMIR Human Factors, we address this challenge through the development of Carer eSupport. This web-based application was designed using a human-centred design (HCD) approach.
The aim of the study was to present how a web-based application can be designed to support ICs’ well-being and preparedness for caregiving through the HCD process. We detail the design and functionalities of the Carer eSupport application and explain how it responds to both the functional and psychological needs of ICs. By integrating HCD with a eudaimonic perspective, our research also intends to provide insights into design approaches that prioritise user well-being in web-based application development.
Additionally, we reported findings from our pilot study and highlighted the initial challenges ICs faced when engaging with the application and the strategies used to overcome these issues.
Moving Beyond Functionality: Designing for Real Needs
We applied the HCD process not only to improve usability, but also to address both the functional and psychological needs of ICs. The design process actively involved ICs, healthcare professionals (HCPs) and researchers across three iterative cycles.
This approach ensured that the application reflects the realities of caregiving. For example, Carer eSupport integrates:
Educational resources to improve preparedness for caregiving tasks
Practical guidance on symptom management and daily care challenges
Self-care support addressing fatigue, stress, and emotional strain
Psychosocial features such as discussion forums and interaction with HCPs
Integrating Well-Being into Design
One of the study’s main contributions is the integration of a eudaimonic perspective into the design of well-being. Rather than focusing solely on short-term satisfaction, the intervention addresses deeper psychological needs, including autonomy, competence, and relatedness. This approach is grounded in eudaimonic well-being, which emphasises living in a way that is meaningful, purposeful, and aligned with one’s values.
Eudaimonic motivation differs from hedonic motivation in that it does not primarily focus on pleasure or comfort. Instead, it emphasises personal growth, meaning, and the realisation of one’s potential. In the context of caregiving, this distinction is particularly relevant. ICs are not only seeking ease of use but are also navigating emotionally demanding situations where purpose, mastery, and connection are central to their experience.
This is where Self-Determination Theory provides a useful theoretical foundation. It proposes that sustained and meaningful motivation is supported by three basic psychological needs:
Autonomy: feeling in control of one’s actions and decisions
Competence: feeling capable and effective in managing tasks
Relatedness: feeling connected to others and supported
The Carer eSupport application operationalises these principles in its design. Autonomy is supported by allowing ICs to navigate content according to their own needs and pace. Competence is strengthened through tailored educational materials and practical guidance. Relatedness is fostered through communication features that connect ICs with healthcare professionals (HCPs) and, where relevant, peers in similar situations.
Importantly, these design choices go beyond improving usability. They support a deeper form of engagement in which ICs are enabled to develop confidence, resilience, and a stronger sense of meaning in their caregiving role. This reflects the core idea of eudaimonic motivation: engagement driven by intrinsic value and personal significance rather than short-term satisfaction.
Evidence from the Pilot Study
The paper further strengthens its contribution by reporting findings from a pilot study, demonstrating the feasibility of the intervention. The results show encouraging engagement, including a strong consent rate and successful user onboarding, confirming readiness for a larger trial.
At the same time, the study does not overlook challenges. ICs faced several barriers when engaging with the application:
Demanding Nature of Caregiving
Due to their caregiving responsibilities, many ICs struggled to find time and energy to engage with the application. During the early phase of diagnosis and treatment, caregiving demands often took priority. In this context, follow-up reminders supported re-engagement once their situations stabilised. One participant noted that although they initially could not log in, a later reminder enabled access when their circumstances improved.
Communication Problems
Some ICs did not receive the initial email with login details or found it in their spam folder, delaying access. To address this, we implemented a multichannel communication strategy using email, SMS, and follow-up phone calls. This improved access speed and strengthened user support.
Login and Authentication Problems
As Carer eSupport is built on Microsoft SharePoint, some users experienced authentication conflicts, particularly when multiple Microsoft accounts were involved. The two-factor authentication process was also perceived as complex. We responded by improving written instructions and developing a step-by-step video tutorial, which reduced login errors and improved independent access.
Access to IT Infrastructure
Some ICs lacked access to suitable devices and relied on smartphones or shared computers, which sometimes limited functionality. As hardware provision was not feasible, we provided extended telephone support. While this enabled eventual access, results suggest that access would be smoother with more consistent use of computers or tablets.
Limited IT Knowledge
Limited digital literacy, particularly among older ICs, also affected onboarding. To address this, we provided structured telephone guidance and clear instructional materials to support the login process and initial use of the system.
Why This Matters
This work provides a compelling example of how HCD can be applied in a structured and meaningful way to design for well-being. It shows that:
Supporting ICs requires more than information; it requires addressing emotional and psychological needs
Iterative, participatory design leads to more relevant and usable interventions
Real-world challenges must be integrated into the design process, not treated as external issues
Most importantly, the study demonstrates that the aim is not only theoretical but achieved in practice: Carer eSupport represents a functioning, user-informed system that supports both preparedness and well-being.
Looking Ahead
While the pilot study confirms feasibility, the ongoing randomised controlled trial will provide further evidence of effectiveness. Nevertheless, this work already offers valuable insights for researchers and designers working at the intersection of HCI, healthcare, and well-being.
It supports a broader shift in digital health design: from systems that primarily deliver information to systems that actively support people in complex and emotionally demanding roles.
Learn more: The full details of our study, including all data and design insights, are published open-access in JMIR Human Factors. Read the paper here: humanfactors.jmir.org/2026/1/e81896.
I mitten av mars höll Jessica Lindblom en keynote-presentation på tvådagarskonferensen Digital teknik i arbetsvardagen – teknikstrul, stress och möjligheter på Arbetets museum i Norrköping. Hon lyfte hur AI, robotar och digitala system påverkar den kognitiva arbetsmiljön och hur gapet mellan ”arbetet som det är tänkt” och ”arbetet som det faktiskt utförs” kan skapa stress, avbrott och ökad mental belastning.
Jessica presenterade aktuella resultat från två olika forskningsprojekt:
AIHURO (Vinnova) – om interaktionen mellan människor och mobila kollaborativa robotar i tillverkningsindustrin. Resultaten visar tydliga skillnader mellan montörer och truckförare i upplevelsen av införandet av robotarna i fabriken, särskilt gällande säkerhet, förutsägbarhet och kognitiv belastning.
AROA (AFA Försäkring) – där hon och kollegorna utvecklar ett ramverk för digitalt arbetsengagemang och hur avancerade system som AI och robotik kan stödja vitalitet, hängivenhet och försjunkenhet i arbetet inom branscherna IT, tåg och lantbruk.
Det blev två trevliga och inspirerande dagar i Norrköping, med många intressanta presentationer – bland annat om algoritmiskt management, framtidens arbetsorganisation och hur olika former av digital teknik och AI (om)formar beslut, kompetenser och arbetsroller i arbetslivet.
Beyond ‘Green Growth’: Why HRI Needs a Post-Growth Perspective
The field of Human-Robot Interaction (HRI) is at a crossroad. As we design more sophisticated robots for healthcare, education, and the home, we are forced to reckon with that our technological progress is deeply intertwined with an economic model that demands perpetual growth—often at a steep cost to our planet.
In our new paper, ‘Post-growth Perspectives in HRI’, recently presented at HRI ’26 in Edinburgh, my co-authors (Mafalda Gamboa, Ilaria Torre, and Birgit Penzenstadler) and I argue that it’s time to look beyond the ‘green growth’ narrative.
The Problem with ‘Business as Usual’
Climate research has pointed out that economic growth is the root cause of environmental devastation, and the most prevalent ‘rescue narrative’ is that productivity growth will recover through new technological breakthroughs. But, this narrative rarely account for the limits to the planetary resources.
Almost all improvements in user interactions increase energy and resource consumption. The carbon emissions from training Large-Language Models is one example, and emissions from sensors and components are also mainly caused by energy consumption. All parts, both software and hardware, of the robots the HRI community are building and buying are, therefore, (somewhat) harmful to our planet.
In parallel, we are seeing a growing graveyard of social robots, where many of the robot platforms that the field has relied on are going bankrupt due to not being able to expand and grow beyond academic buyers. We argue that HRI cannot solely rely on the unstable market of social robots and the market mechanisms that the field is deeply tied to, such as software-as-a-service and robot-as-a-service, which by many large companies is often driven by robotwashing. To tackle this, we point to self-built and low-scale robots, even though they are not exempt from the material dependencies of the broader growth paradigm.
Introducing Post-growth for HRI
Our paper introduces a post-growth perspective to HRI, by inviting the community to shift the focus from quantitative growth to qualitative development. We reuse the metaphor of the snail (broadly used by the degrowth and slow movements) and add a new ring on the snail’s shell—a layer of automation, artificial intelligence, and robotics, built upon already complex structures. While intended to enhance efficiency and productivity, this continual layering risks creating fragility, dependence, and ecological or social imbalance. The metaphor thus warns of the limits of accumulation and invites reflection on how to achieve balance for HRI.
To help researchers and practitioners navigate this shift, we propose three actionable steps:
Ask ‘Question Zero’: Inspired by the question zero for the use of AI: ‘why AI?’, we, in our version, ask ‘why robots?’. Asking this forces criticality and more detailed motivation for why a robot is needed and for whom. Question zero can also inform research that examines real-world applications in need of critique. Beyond refusal, the question seeks to find cases in need of assessment and possible exnovation.
The Post-growth Manifesto: We suggest a manifesto (see below), that comes with a glossary. Here, we list a number of conventional understandings of robots in HRI based on our reading of the field, in opposition to a list of important notions within post-growth for HRI. It can be a tool to locate, discuss, and negotiate ongoing research, and in doing so, we encourage leaning into the B side to support a post-growth orientation.
Sustainability Statements: Just as ethical and positionality statements have become standard, we propose a template for sustainability statements. These encourage researchers to reflect on the environmental, social, and economic impacts of their work.
Why This Matters
As HRI researchers, we have the power to define what the ‘future’ looks like. By embracing post-growth, we aren’t ‘anti-technology’. Instead, we are advocating for a future where robotics serves humanity and the biosphere without being tied to the destructive cycle of endless extraction.
We hope this paper serves as a call to reflexivity and criticality. It’s time to consider not just how we build and deploy robots, but for whom and at what cost.
Citation:Sofia Thunberg, Mafalda Gamboa, Ilaria Torre, and Birgit Penzenstadler. 2026. Post-growth Perspectives in HRI. In Proceedings of the 21st ACM/IEEE International Conference on Human-Robot Interaction (HRI ’26). Association for Computing Machinery, New York, NY, USA, 1388–1398.
In our research group, we study the relationships and dynamics of Human, Technology, and Organisation (HTO) to create knowledge that supports sustainable development and utilization of ICT.
