The Commonplace
Home Papers Evidence Explore Trends Syntheses Digests About 🎲 Workforce Futures
← Papers
Direction, evidence grade, and study type are AI-generated labels (gpt-5-mini), not human-verified. Syntheses are LLM-written. "Tensions" are machine-detected candidates, not confirmed contradictions. A research-acceleration tool, not peer review. How this is built →

Software engineers treat AI as an intellectual collaborator rather than a social teammate, expecting fewer socio-emotional attributes from models and instead valuing functional capabilities like contextual adaptation and responsibility negotiation; the authors argue teams should design for 'functional equivalents' rather than mimic human socio-emotional intelligence.

Bridging the Socio-Emotional Gap: The Functional Dimension of Human-AI Collaboration for Software Engineering
Lekshmi Murali Rani, Richard Berntsson-Svensson, R. Feldt · Fetched April 04, 2026 · arXiv.org
semantic_scholar descriptive low evidence 7/10 relevance Full text usable extracted full text DOI Source PDF

Structured author observations

Linked only from stored provider relations; the raw author line above is never matched by name.

OpenAlex

Latest observation:

  1. Rani, Lekshmi Murali provider ID
  2. Svensson, Richard Berntsson provider ID
  3. Feldt, Robert provider ID

Semantic Scholar

Latest observation:

  1. Lekshmi Murali Rani provider ID
  2. Richard Berntsson Svensson provider ID
  3. R. Feldt provider ID
Software practitioners view AI as intellectual teammates rather than social partners and prioritize functional capabilities—contextual intelligence, adaptive learning, collaborative intelligence—over human-like socio-emotional traits for effective collaboration.

Citation observations

Cumulative provider counts captured on specific dates; providers are never combined.

As GenAI models are adopted to support software engineers and their development teams, understanding effective human-AI collaboration (HAIC) is increasingly important. Socio-emotional intelligence (SEI) enhances collaboration among human teammates, but its role in HAIC remains unclear. Current AI systems lack SEI capabilities that humans bring to teamwork, creating a potential gap in collaborative dynamics. In this study, we investigate how software practitioners perceive the socio-emotional gap in HAIC and what capabilities AI systems require for effective collaboration. Through semi-structured interviews with 10 practitioners, we examine how they think about collaborating with human versus AI teammates, focusing on their SEI expectations and the AI capabilities they envision. Results indicate that practitioners currently view AI models as intellectual teammates rather than social partners and expect fewer SEI attributes from them than from human teammates. However, they see the socio-emotional gap not as AIs failure to exhibit SEI traits, but as a functional gap in collaborative capabilities (AIs inability to negotiate responsibilities, adapt contextually, or maintain sustained partnerships). We introduce the concept of functional equivalents: technical capabilities (internal cognition, contextual intelligence, adaptive learning, and collaborative intelligence) that achieve collaborative outcomes comparable to human SEI attributes. Our findings suggest that effective collaboration with AI for SE tasks may benefit from functional design rather than replicating human SEI traits for SE tasks, thereby redefining collaboration as functional alignment.

Summary

Main Finding

Software practitioners do not primarily expect human-like socio-emotional intelligence (SEI) from AI teammates. Instead, they see the “socio‑emotional gap” as a functional collaboration gap: AI lacks capabilities that achieve the same collaborative outcomes as human SEI (e.g., negotiating responsibilities, contextual adaptation, continuity). The authors propose a “functional equivalents” framework — four technical capabilities (internal cognition, contextual intelligence, adaptive learning, collaborative intelligence) that can substitute for human SEI in software engineering (SE) tasks without replicating human emotions.

Key Points

  • Role perception

    • Practitioners treat current generative-AI tools as intellectual/cognitive teammates (code generation, debugging, knowledge search), not social partners.
    • They expect far fewer SEI attributes (trust-building, empathy, rapport) from AI than from human teammates.

  • Reconceptualizing the gap

    • The socio-emotional gap is framed functionally: AI’s missing abilities to sustain partnership, negotiate and share responsibilities, preserve context across interactions, and adapt to team dynamics.
    • Practitioners prioritize functional outcomes (reliability, context continuity, explainability) over anthropomorphic SEI.

  • Functional equivalents framework (derived from mapping ESCI-U SEI traits to technical features)

    • Internal cognition: internal state representation, self-monitoring, explicit uncertainty and capability signalling.
    • Contextual intelligence: sustained context/long-term memory, project and domain awareness, awareness of team conventions and constraints.
    • Adaptive learning: rapid online adaptation to user/workflow preferences, learning from corrections, incremental improvement within a project.
    • Collaborative intelligence: negotiation of responsibilities, task handoffs, multi-agent coordination protocols, explainable reasoning for joint decisions.

  • Empirical grounding and validation

    • Qualitative study with semi-structured interviews (Phase 1: 6 participants; Phase 2: 4 additional participants) of SE practitioners experienced with tools like Copilot, ChatGPT, Claude.
    • Thematic analysis (Braun & Clarke) produced seven themes; functional equivalents framework was refined and validated in Phase 2 interviews.

  • Limits called out by authors

    • Small, purposive sample (n=10) of technically proficient practitioners; results may not generalize to non-technical users or other sectors.
    • Practitioners’ views reflect current AI maturity; preferences might shift as capabilities evolve.

Data & Methods

  • Research design: exploratory qualitative study using semi-structured interviews guided by ESCI-U (emotional, social, and cognitive intelligence vocabulary).
  • Participants: 10 software practitioners (roles: software engineers, ML engineers, researchers, consultants) with 3–11 years SE experience and 1–3 years AI tool experience. Recruited purposively via LinkedIn and networks.
  • Data collection: Two phases. Phase 1 (6 interviews) used ESCI-informed interview protocol to elicit perceptions of SEI in human vs. AI teammates and collaboration challenges. Phase 2 (4 interviews) validated the drafted functional equivalents framework with use-case scenarios.
  • Analysis: Thematic analysis following Braun & Clarke’s six-phase approach; open and axial coding to derive themes; mapped ESCI traits to technical capabilities to produce the framework. Thematic saturation monitored; Phase 2 used to validate and refine themes.
  • Data availability: Interview instruments referenced (Zenodo). Limitations: small sample, practical rather than statistical saturation, possible subjectivity in interpretations.

Implications for AI Economics

  • Task allocation and complementarity

    • Practitioners view AI as a cognitive/computational complement rather than a social substitute. Economic models should treat AI adoption in SE as shifting task bundles (more cognitive/analytical automation) while leaving socio-emotional coordination tasks differently allocated.
    • Functional capabilities that substitute for coordination (e.g., automated handoffs, context persistence) alter the complementarity between AI and workers, affecting demand for specific human skills (less routine cognitive work; persistent need for high‑level social coordination, domain judgment).

  • Productivity and value capture

    • Value from AI in SE may come less from simulated empathy/SEI and more from measurable functional improvements (reduced context-switching, fewer integration errors, faster onboarding, lower coordination overhead). Pricing and ROI assessments should prioritize metrics tied to these functional gains.
    • Vendors and firms that invest in the four functional equivalents are likely to capture greater economic value than those investing primarily in anthropomorphic SEI affordances.

  • Labor market and wage effects

    • Demand for workers with skills in supervising/adapting AI (meta-cognitive and strategic roles) may increase. Conversely, roles focused on routine coordination or persistent contextual memory (e.g., certain integration tasks) may see automation pressure if AI achieves functional equivalents.
    • Wage polarization is possible: premium for engineers who can orchestrate human-AI workflows and for managers who translate sociotechnical requirements into machine-ready specifications.

  • Product design, adoption, and procurement

    • Firms procuring AI tools should prioritize functional features (context persistence, explainability, negotiation APIs) for SE teams. This changes procurement criteria, contracting language, and vendor competition (differentiation by functional capabilities).
    • Pricing models could shift from per-query to per-project or per-context models that value sustained context and long-term adaptation (subscriptions with project memory or team-level licensing).

  • Measuring impact — empirical research agenda for economists

    • New outcome metrics: time-to-resolution for bugs, handoff error rates, context-reconstruction time, frequency of human overrides, and team coordination costs.
    • Suggested empirical strategies:
      • Randomized controlled trials (A/B) within firms to test features implementing functional equivalents (e.g., persistent project memory vs. stateless assistants).
      • Difference-in-differences using staggered firm/ team adoption to estimate productivity effects and wage impacts.
      • Task-level time-use studies and matched employer-employee panel data to observe allocation shifts and wage trajectories.
      • Field experiments that vary the AI’s transparency (self-monitoring signals, uncertainty reports) to measure trust, reliance, and decision quality.
    • Causal identification: exploit rollouts at project/team granularity and instrument for feature exposure (e.g., vendor update schedules).

  • Policy and investment implications

    • Public and private R&D funding should prioritize functional collaboration features (context continuity, explainable negotiation) that yield measurable productivity gains, not just anthropomorphic interfaces.
    • Workforce policy: invest in retraining for supervisory and orchestration roles, and create standards for measuring AI-assisted productivity in organizational settings.
    • Procurement policy in governments and large organizations should require evaluation of functional equivalents (project memory, responsibility negotiation) for AI tools used in collaborative knowledge work.

  • Market structure and competition

    • If functional equivalents are harder to copy (require long-term project memory and team-specific adaptation), incumbent platforms with long interaction histories may gain lock-in advantages, affecting competition and market concentration in AI tooling for enterprises.
    • Economists should study how "context stickiness" (value of preserved context) affects switching costs and bar entry.

  • Cautions for economists

    • Findings are from technically proficient SE practitioners; non-technical sectors may place higher value on anthropomorphic SEI, so cross-sector heterogeneity is expected.
    • As AI evolves, the boundary between functional and socio-emotional may blur; dynamic models are needed.

Suggested immediate empirical questions for AI economists - How much of measured productivity gains from AI in software teams is explained by functional equivalents (context continuity, fewer handoffs) versus raw code generation? - Do teams that adopt AI tools with stronger contextual/adaptive features experience lower coordination costs and different wage trajectories than teams using stateless assistants? - What is the effect of explicit uncertainty/ability signaling by AI on human reliance, error rates, and blame attribution in firms?

Overall, this paper reframes socio-emotional deficits of AI in SE as a design problem of missing functional collaboration capabilities. For AI economists, that suggests shifting measurement and policy attention toward the economic effects of those functional equivalents (productivity, labor demand, market structure, procurement), and designing empirical work around concrete operational metrics rather than anthropomorphism alone.

Assessment

Paper Typedescriptive Evidence Strengthlow — Findings are based on n=10 semi-structured interviews and reflect practitioners' perceptions rather than measured behavioral or productivity outcomes; small, non-representative sample and qualitative design limit external validity and preclude causal claims. Methods Rigormedium — Use of semi-structured interviews and thematic analysis is appropriate for exploratory work and the paper articulates a clear conceptual contribution (functional equivalents), but the small sample, limited triangulation, and reliance on self-report constrain methodological robustness. SampleTen software engineering practitioners were interviewed in semi-structured sessions about their expectations and experiences collaborating with AI vs. human teammates; data are qualitative, self-reported, and analyzed thematically (no randomized or behavioral measurement). Themeshuman_ai_collab org_design skills_training GeneralizabilitySmall sample size (n=10) limits statistical and population-level generalization, Potential selection and volunteer bias toward practitioners with interest or experience in AI tools, Findings specific to software engineering contexts and may not transfer to other occupations, Lack of demographic, company-size, geographic, and domain diversity reduces external validity, Perception-based data may not predict actual collaboration outcomes or productivity

Claims (10)

ClaimDirectionOutcomeConfidence & EvidenceDetails
Socio-emotional intelligence (SEI) enhances collaboration among human teammates. Team Performance positive quality of collaboration among human teammates
Reading fidelity high
Study strength medium
not reported
0.18
Current AI systems lack SEI capabilities that humans bring to teamwork, creating a potential gap in collaborative dynamics. Team Performance negative presence of SEI capabilities in AI systems (vs. humans)
Reading fidelity high
Study strength low
not reported
0.09
This study uses semi-structured interviews with 10 practitioners to examine perceptions of collaborating with human versus AI teammates. Other null_result methodological description (data collection approach)
Reading fidelity high
Study strength high
n=10
0.3
Practitioners currently view AI models as intellectual teammates rather than social partners and expect fewer SEI attributes from them than from human teammates. Team Performance negative practitioners' expectations of SEI attributes in AI versus human teammates
Reading fidelity high
Study strength medium
n=10
0.18
Practitioners see the socio-emotional gap not as AI's failure to exhibit SEI traits, but as a functional gap in collaborative capabilities. Task Allocation mixed framing of the AI–human socio-emotional gap (functional vs. emotional)
Reading fidelity high
Study strength medium
n=10
0.18
Practitioners identified specific functional deficiencies in AI: inability to negotiate responsibilities. Task Allocation negative AI capability to negotiate responsibilities in teamwork
Reading fidelity high
Study strength medium
n=10
0.18
Practitioners identified specific functional deficiencies in AI: inability to adapt contextually. Organizational Efficiency negative AI capability for contextual adaptation in collaborative work
Reading fidelity high
Study strength medium
n=10
0.18
Practitioners identified specific functional deficiencies in AI: inability to maintain sustained partnerships. Team Performance negative AI capability to maintain sustained collaborative partnerships
Reading fidelity high
Study strength medium
n=10
0.18
The authors introduce the concept of 'functional equivalents': technical capabilities (internal cognition, contextual intelligence, adaptive learning, and collaborative intelligence) that achieve collaborative outcomes comparable to human SEI attributes. Team Performance positive ability of technical capabilities to achieve collaborative outcomes comparable to human SEI
Reading fidelity high
Study strength speculative
not reported
0.03
Effective collaboration with AI for software engineering (SE) tasks may benefit from functional design rather than replicating human SEI traits, thereby redefining collaboration as functional alignment. Team Performance positive effectiveness of human-AI collaboration in SE tasks
Reading fidelity high
Study strength speculative
n=10
0.03

Notes