Finding: This systematic scoping review maps AI applications for neurodiverse learners (ASD, ADHD, dyslexia) across five domains, finding that AI shows promise for personalized learning and assistive technologies but faces significant risks from bias, privacy concerns, and insufficient inclusion of neurodivergent voices in design.
Why it matters: This directly informs HCI design principles for inclusive AI systems and highlights critical human-centered design considerations when developing AI tools for diverse learners.
Method: PRISMA-ScR systematic scoping review methodology covering 2017-2025 literature across education, computing, and interdisciplinary databases.
Jargon: PRISMA-ScR = Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews; neuroadaptive tools = technologies that adapt to users' cognitive states and neural patterns.
Finding: The book introduces "synthetic intelligence" as a harmonious blend of human and artificial intelligence in education, providing frameworks for meaningful human-AI collaboration and co-learning experiences that go beyond traditional educational approaches.
Why it matters: This directly addresses AI's impact on learning sciences and human-AI collaboration, offering practical frameworks for educators navigating the integration of AI tools in educational contexts.
Method: Not applicable - this is a theoretical framework book.
Jargon: Synthetic intelligence - the harmonious integration of human cognitive abilities with AI capabilities for collaborative learning and problem-solving.
Finding: Eight large language models were evaluated using human cognitive screening tests (MMSE, CDT, MoCA), with Claude Sonnet 4.5 performing best but still failing to achieve maximum scores, and all models showing poor performance on visual tasks like clock drawing compared to language tasks.
Why it matters: This reveals important limitations in AI cognitive capabilities that are relevant for understanding how AI can effectively collaborate with humans and where human-AI task allocation should be optimized.
Method: Applied three standardized cognitive assessment tools designed for humans to eight different LLMs, with each test performed three times per model.
Jargon: MMSE (Mini Mental Status Examination), CDT (Clock Drawing Test), MoCA (Montreal Cognitive Assessment) - standardized tests used to screen for cognitive impairment in humans; modality gap - performance differences between visual and language processing tasks.
⚠️ No control group mentioned | Single study (not replicated) | Limited statistical detail in abstract