2026 ASEE Annual Conference & Exposition

In this paper, we present the design of an NSF Grantees Poster synthesizing two years of implementation and evaluation findings from the Preparing Autistic Students for the AI Workforce (PAS4AI) project. The poster shows how our program evolved from technical skills instruction to an integrated approach to addressing the critical gap between technical competence and workforce readiness for autistic community college students. This evolution responds directly to Year 1 external evaluation findings that identified persistent interview anxiety, uneven instructor preparation, and the need for explicit social-emotional learning (SEL) support as barriers to student success.

We use a three-panel structure to communicate our core design pattern. The left panel establishes the problem space, presenting evidence from program data and literature on why SEL functions as the essential link between technical skill and workforce success for autistic learners. We foreground PAS4AI Year 1 evaluation feedback, including IRB and recruitment constraints, the implementation of daily feedback loops, and explicit recommendations to strengthen SEL and mentor training.

The center panel visualizes the SEL-integrated design, mapping the linked interventions. First, we present SEL as teachable microskills rather than abstract concepts, operationalized through discrete, coachable behaviors such as recognizing overwhelmedness and selecting communication strategies. Next, we detail the interview process we used with students: repeated, low-stakes mock interviews using neurodiversity-friendly rubrics, transparent LLM rehearsal protocols with ethical boundaries, and structured guidance on disclosure decisions.

The right panel documents evidence and forward-looking insights from Year 2 qualitative analysis, including the deliberate pivot toward self-efficacy, identification of interviews as critical barriers, emerging roles for LLMs as non-judgmental rehearsal partners, and the rationale for entrepreneurship pathways offering workplace sovereignty for learners who struggle in hierarchical environments. We propose a unified evaluation framework linking SEL gains, interview fluency, and portfolio performance to workforce outcomes.

The poster addresses a key Year 1 finding: instructor variability in implementing inclusive practices. To ensure consistent, effective teaching in PAS4AI, we developed a scenario-based train-the-trainer model covering SEL microskills coaching, ethical LLM rehearsal boundaries, and non-verbal collaboration scaffolds.

This work contributes a replicable framework and open educational resources that other AI and computing programs can adapt to serve neurodivergent learners. The poster extends our 2025 ASEE findings on instructor experience and social-emotional demands in AI education while responding systematically to external evaluator recommendations.