Jump to content

Draft:Socio-cognitive engineering

From Wikipedia, the free encyclopedia
  • Comment: Has AI-indicative tone and structuring issues. AI tends to make overly vague statements without any detail. Please expand those bullet points into proper sentences and reduce generalizations and vagueness. Ca talk to me! 09:40, 7 June 2025 (UTC)

Socio-Cognitive Engineering (SCE) is a methodological framework for developing Artificial Intelligence (AI)-integrated technologies that are intended to complement and augment human activity in ways that align with human motives and values. It is particularly focused on support systems for critical human tasks in demanding, high-responsibility situations, such as emergency response, chronic health management (e.g., diabetes self-care), support for independent living among the elderly, military operations, and long-duration space missions. SCE employs an iterative, ongoing design-and-test process within the broader vision of Hybrid Intelligence (HI)—collaborative systems that integrate human and artificial capabilities[1][2]. Drawing on knowledge, models, and methods from multiple disciplines—including human-centered design, cognitive science, requirement engineering, human factors, and value sensitive design.—SCE promotes a transdisciplinary approach. A distinguishing feature of the methodology is the explicit articulation of design rationales, including theoretically and empirically grounded claims about functional performance, human experience and value alignment. SCE is a developing approach that requires progress on its adoption, and sharing of best practices:

Origins and Theoretical Foundations

[edit]

The term socio-cognitive[3] reflects the mutual influence between individual cognitive processes and social dynamics, an important aspect of human–computer interaction that most often takes place in a social context (cf., social cognition and situated cognition). SCE evolved from work in Cognitive Systems Engineering (CSE), which emerged in the 1980s to improve human performance in complex and high-risk domains such as aviation, nuclear power, and process control.[4]

The first notion of Socio-Cognitive Engineering as a design methodology was presented by Sharples et al. (2002), who proposed a stakeholder-centered process that integrates cognitive models, usability evaluation, and task analysis in human-centered system design.[5]

Subsequently, Neerincx and Lindenberg (2008) introduced Situated Cognitive Engineering (SCE), extending CSE analysis methods with scenario-based design, cognitive task modeling and empirical evaluation techniques.[6] Their elaboration of the methodology was based on large research programs in two high-stake domains: the development of astronaut support for manned space missions, and support development for ship control and mission command teams in naval missions. In both domains, operational procedure design & training, technology development and interaction design were performed by rather separated communities, resulting in diverse misalignments (like underutilization of technical capabilities, and incoherence between procedures and user interfaces). Building on available methods, as described above, to address these needs for a practical coherent research and development of intelligent work support, the SCE development started. SCE itself was developed through an iterative process, being refined based on practical experience of real-world cases and incorporating advances from the field of cognitive engineering.

The evolution into Socio-Cognitive Engineering was further formalized in Neerincx et al. (2019) through its application in health robotics with the example of a robotic-and-agentic partner with dedicated dashboards to support children with type 1 diabetes and their caregivers [7] In the foundation phase, stakeholder needs and values were acquired (e.g. autonomy), relevant psychological theories identified (e.g., Self-Determination Theory), and technological solutions chosen (e.g. eXplainable AI), forming a knowledge, requirements and technology base. The specification phase defined partnership functions with claims (e.g., joint goal setting, feedback exchange, and experience sharing to improve child's knowledge and self-efficacy). The evaluation phase involved iterative user studies to test these claims in real-life settings (e.g., camps and clinics) to assess the robot’s usability, engagement, and support for autonomy, competence, and relatedness. Generic knowledge and design solutions have been incorporated in a "Partnership" ontology for re-use.

Methodological Framework

[edit]

As the example of diabetes management support shows (see above), the Socio-Cognitive Engineering (SCE) is structured as an iterative, four-phase methodology. It takes a transdisciplinary approach to address both the diversity of theories and models that can enhance the design and evaluation and the domain and context-dependent requirements and effects. There are four established methods with key principles that SCE incorporated: First, Norman's (1986) work on user-centered design and cognitive artifacts[8] emphasizes the importance of design, usability, and end-user evaluation for appropriate adoption of new interactive technology (addressing the psychological aspects, like end-user vs. software-engineer mental models). The foundation phase includes an analysis of op the operational demands (user needs) and the relevant cognitive theories (like Self-Determination Theory for diabetes management, and Workload Theory for ship operator support). Second, Vicente's (1999) ecological interface design[9] added a socio-technical perspective to develop displays that accommodate for expert knowledge to control complex dynamic processes. Correspondingly, the foundation phase (as part of the operational demands) includes focused work domain analyses and representations (like the ship control resource states to display). Third, Carroll's (2000) on scenario-based design[10] approach showed how the construction of problem and design scenarios can facilitate the iterative development process to address task-artefact cycle (i.e., a new artefact brings about new tasks that set new requirements for a new artefact, ...), which is includes in the work domain analyses. Fourth, Friedman et al. work (2006) on value sensitive design[11] provide sound value analyses method that can facilitate human-technology value alignment, which is being included in the operational demand analyses in the form of value stories.

1. Foundation

[edit]

The first phase identifies and synthesizes core requirements by analyzing the operational context, cognitive demands, and technological constraints. This process is structured around three main perspectives:

  1. Operational Demands: Analyses are based on in-depth investigations of the work environment and tasks. Techniques such as ethnographic fieldwork, work domain analysis [12], and goal-directed task modeling [13] are used to understand the structure and dynamics of user activities within the specific domain.
  2. Human Factors and Cognitive Theories: This involves applying theoretical models to guide design assumptions. Examples include workload theory[14] to estimate cognitive-affective task load and mental effort, mental models[15] to understand user expectations and behaviors, and situation awareness frameworks [16] to support real-time comprehension and decision making.
  3. Technological Constraints and Opportunities: The analysis includes examining the capabilities and limitations of AI technologies, interaction modalities (e.g., voice, gesture, mixed reality), and sensor-actuator systems. These are assessed for their suitability within the target environment and tasks.

2. Design Specification

[edit]

This phase translates the foundational insights into concrete envisioned system behavior and interaction strategies (e.g. on situation awareness displays, task planning & allocation, warnings & advises, feedback & explanations). The core of this specification consists of two components that are being worked in combination. The first component concerns the generation of Use Cases, as common in software engineering but with an emphasize on the human-technology interaction. Detailed narratives or scenarios are developed to describe system functionalities and user interactions. These use cases are grounded in the operational context and aim to cover both routine and edge-case situations[10]. Second, the Claims Analysis is being worked out, i.e., each design choice is linked to specific outcomes through testable claims. These claims articulate assumed or evidenced cause–effect relationships between system features and their anticipated effects on concrete measures, e.g. on performance, workload, learning, trust, situation awareness, safety and well-being. [17]

3. Evaluation

[edit]

The evaluation is embedded throughout the development cycle and relies on a combination of qualitative and quantitative methods to test claims, identify value mis-alignments, and inform design iterations. It is integral and recursive, informed by principles of mixed methods research. Also here, two central components are distinguished. First, prototypes are built that may range from simulations, early mock-ups to functional systems (it can also be combinations of "real technology" and simulations). Functions may be partially included in a prototype to test specific claims. In general, prototypes are used in controlled settings or in the field to generate user feedback and empirical data. Second, the evaluation methods concern multiple empirical approaches. Quantitative methods include performance metrics and standardized usability assessments. Qualitative methods such as semi-structured interviews, think-aloud protocols, and field observations help interpret user experience. Evaluation is conducted through iterative cycles that may include laboratory experiments, cognitive walkthroughs, ethnographic studies, and field trials. Overall, the techniques are combined to assess claims and guide refinements. The combination of methods (e.g., controlled experiments, cognitive walkthroughs, ethnographic observation, and field trials) allows for triangulation of results.[18][19]

4. Abstraction for Coherence and Reuse

[edit]

To support generalization and knowledge reuse, this phase focuses on formalizing insights and making them applicable across domains. The knowledge abstraction concerns three models.

  1. Values: Ethical and societal considerations are translated into actionable design criteria and evaluation benchmarks. These are not treated as abstract principles but as operational requirements derived from stakeholder input and value-sensitive design processes. Note that scenarios are used to facilitate reflection, deliberation and modeling of the values (e.g., to be included in requirements engineering processes); these context-dependencies are included in the value models..[11]
  2. Design Patterns: Common solutions to recurring interaction and interface challenges are abstracted into reusable patterns. These patterns are documented with guidance on their applicability (context-dependency), limitations, and associated empirical findings. Furthermore, the related requirements to their part of the work are explicated, both for the AI and the human[20]
  3. Ontologies: Domain-specific ontologies are developed to represent task structures, system functions, mental models and user goals. These formal models enable consistency in design, support model-driven development, and facilitate interoperability across systems and components[21]

Applications

[edit]

SCE has been applied in several domains where collaboration between humans and intelligent agents is essential (human–AI collaboration). The methodology has supported system design and evaluation in the following areas.

In the domain of Health and Wellbeing, for digital health interventions and social robotics, SCE has been used to guide the design of e-coaching systems, serious games, and social robots (companion robots). These systems support long-term behavior change and chronic disease management by aligning system functions with user goals, values, and cognitive capacities.[7]

Further, SCE informed the development of intelligent cognitive agents to assist astronauts with planning, monitoring, and anomaly handling in the domain of Space Exploration. For instance, in the European Space Agency’s MECA project, the methodology helped define the agent’s interaction capabilities and ensure that its reasoning processes were transparent and aligned with human expectations.[22]

In the domain of Traffic Management (e.g. rail traffic): SCE has been applied to the design of decision-support systems that help distribute cognitive workload among human operators (e.g. train dispatchers) and AI agents. This includes techniques to model team coordination and dynamically adjust automation levels to avoid overload or underload.[14]

As another example, now for Disaster Response, SCE supported the design of robotic platforms and coordination tools used in emergency response scenarios. SCE helped identify critical information needs, evaluate situation awareness support under high workload, and ensure that air and ground vehicles (robots) remained under meaningful human control.[23]

In Defense and Security applications, SCE has been used to support human–machine teaming for tasks such as situational awareness, threat assessment, and tactical decision-making. The methodology guided the design of actionable explanations to improve trust calibration and design of a work agreements model to attune information processes and responsibilities to the momentary context .[24]

Discussion

[edit]

Socio-Cognitive Engineering (SCE) offers a methodology for the design of interactive technology involving human–AI collaboration, particularly for domains characterized by complexity, uncertainty, volatility and high stake. Its main contribution lies in the structured approach to establish an evolving, concise and coherent knowledge base from an operational (domain), human factors and technical perspective. With these three perspectives, adaptive technology can be designed to complement and augment humans situated cognition. Design choices and evaluations are guided by use case–based claim analyses, which explicitly articulate the underlying design rationale. Dependencies specific to each use case are made transparent, helping to identify components that are generalizable and potentially reusable across domains.

The methodology is based on the notion that both human and technology evolve over time, mutually influencing each other. SCE has been developed and applied in iterative, mixed-method development cycles, combining theoretical analysis with practical feedback. Its emphasis on participatory design, prototyping and empirical validation supports concretization of stakeholders' values and early identification of design shortcomings. However, specific technology advancements take place in a broader technology environment with various dependencies. So far, SCE has not been covering this aspect and been applied in or after implementation processes. In current projects in the health care and crises management domains, we are working towards to include this implementation phase.

Another area for further development involves the scalability of design patterns and ontologies across domains. While abstraction is a key feature of the methodology, generalizing knowledge without oversimplifying context-specific constraints remains an ongoing tension. Libraries of collaboration and design patterns are under development, but still have to prove their value for the research and development community in general. SCE remains a developing methodology that benefits from continued refinement through diverse applications. Its future effectiveness depends on broader adoption across domains, as well as the creation of open platforms for sharing experiences, validating claims, and reusing design rationales, methods, and tools.

See Also

[edit]

References

[edit]
  1. ^ Dellermann, Dominik; Ebel, Philipp; Söllner, Matthias; Leimeister, Jan Marco (2019-10-01). "Hybrid Intelligence". Business & Information Systems Engineering. 61 (5): 637–643. arXiv:2105.00691. doi:10.1007/s12599-019-00595-2. ISSN 1867-0202.
  2. ^ Akata, Zeynep; Balliet, Dan; De Rijke, Maarten; Dignum, Frank; Dignum, Virginia; Eiben, Guszti; Fokkens, Antske; Grossi, Davide; Hindriks, Koen; Hoos, Holger; Hung, Hayley; Jonker, Catholijn; Monz, Christof; Neerincx, Mark; Oliehoek, Frans; Prakken, Henry; Schlobach, Stefan; Van Der Gaag, Linda; Van Harmelen, Frank; Van Hoof, Herke; Van Riemsdijk, Birna; Van Wynsberghe, Aimee; Verbrugge, Rineke; Verheij, Bart; Vossen, Piek; Welling, Max (2020). "A Research Agenda for Hybrid Intelligence: Augmenting Human Intellect With Collaborative, Adaptive, Responsible, and Explainable Artificial Intelligence". Computer. 53 (8): 18–28. doi:10.1109/MC.2020.2996587. Retrieved 2025-06-27.
  3. ^ Sociocognitive. Wikipedia.
  4. ^ Rasmussen, Pejtersen, & Goodstein, 1994; Hollnagel & Woods, 2005.
  5. ^ Sharples, M., Jeffries, H., du Boulay, B., Teather, D., & du Boulay, G. (2002). Socio-cognitive engineering: A methodology for the design of human-centred technology. European Journal of Operational Research, 136(2), 310–323.
  6. ^ Neerincx, M. A., & Lindenberg, J. (2008). Situated cognitive engineering for complex task environments. Ashgate Publishing.
  7. ^ a b Neerincx, M. A., Van Vught, W., Blanson Henkemans, O., Oleari, E., Broekens, J., Peters, R., ... & Bierman, B. (2019). Socio-cognitive engineering of a robotic partner for child's diabetes self-management. Frontiers in Robotics and AI, 6, 118. https://doi.org/10.3389/frobt.2019.00118
  8. ^ Norman, D. A. (1986). The design of everyday things. Basic Books.
  9. ^ Vicente, K. J. (1999). Cognitive Work Analysis: Toward safe, productive, and healthy computer-based work. CRC Press.
  10. ^ a b Carroll, J. M. (2000). Making use: scenario-based design of human–computer interactions. MIT Press.
  11. ^ a b Friedman, B., Kahn, P. H., & Borning, A. (2006). Value sensitive design and information systems. In Human-computer interaction and management information systems: Foundations, 348–372.
  12. ^ Vicente, K. J. (1999). Cognitive work analysis: Toward safe, productive, and healthy computer-based work. CRC press.Endsley (1995)
  13. ^ Neerincx, M. A. (2003). Cognitive task load analysis: Allocating tasks and designing support. In D. A. Schraagen, S. F. Chipman, & V. L. Shalin (Eds.), Handbook of cognitive task design (pp. 283–306). CRC Press.
  14. ^ a b Harbers, M., & Neerincx, M. A. (2017). Value sensitive design of a virtual assistant for workload harmonization in teams. Cognition, Technology & Work, 19(2–3), 329–343.
  15. ^ Johnson-Laird, P. N. (1983). Mental models: Towards a cognitive science of language, inference, and consciousness. Harvard University Press.
  16. ^ Endsley, M. R. (1995). Toward a theory of situation awareness in dynamic systems. Human factors, 37(1), 32-64.
  17. ^ McCrickard, D. S., Catrambone, R., Chewar, C. M., & Stasko, J. T. (2003). Establishing tradeoffs that leverage attention for utility: Empirically evaluating information display in notification systems. International Journal of Human-Computer Studies, 58(5), 547–582.
  18. ^ Creswell, J. W., & Plano Clark, V. L. (2017). Designing and conducting mixed methods research (3rd ed.). Sage Publications.
  19. ^ Taherdoost, H. (2022). What are different research approaches? Comprehensive review of qualitative, quantitative, and mixed method research, their applications, types, and limitations. Journal of Management Science & Engineering Research, 5(1), 53–63.
  20. ^ Van Zoelen, E., Mioch, T., Tajaddini, M., Fleiner, C., Tsaneva, S., Camin, P., ... & Neerincx, M. A. (2023). Developing team design patterns for hybrid intelligence systems. In HHAI 2023: Augmenting Human Intellect (pp. 3-16). IOS Press.
  21. ^ Rijgersberg-Peters, R., van Vught, W., Broekens, J., & Neerincx, M. A. (2023). Goal Ontology for Personalized Learning and Its Implementation in Child's Health Self-Management Support. IEEE Transactions on Learning Technologies, 17, 903-918.
  22. ^ Neerincx, M. A. (2011). Situated cognitive engineering for crew support in space. Personal and Ubiquitous Computing, 15(5), 445–456.
  23. ^ Kruijff-Korbayová, I., Colas, F., Gianni, M., Pirri, F., de Greeff, J., Hindriks, K., ... & Worst, R. (2015). Tradr project: Long-term human-robot teaming for robot assisted disaster response. KI-Künstliche Intelligenz, 29, 193–201.
  24. ^ De Greef, T. E. G., Henryk, F. A., & Neerincx, M. A. (2010). Adaptive automation based on an object-oriented task model: Implementation and evaluation in a realistic C2 environment. Journal of Cognitive Engineering and Decision Making, 4(2), 152–182.
Retrieved from "https://en.wikipedia.org/w/index.php?title=Draft:Socio-cognitive_engineering&oldid=1303334912"