OECD: AI For Assistive Technology And Labour (Report and Repository)

Following our previous cooperation, we contributed to the major work driven by OECD - report and repository (1), exploring AI-driven assistive technologies, its adoption, challenges and implications for labour and workforce. Building on interviews with more than 70 stakeholders and analysis of 142 AI-powered solutions, this report explores the potential of AI to foster employment for people with disability, accounting for both the possibilities of AI-powered solutions and the risks attached to the increased use of AI for people with disability. It also identifies obstacles hindering the use of AI and discusses what governments could do to avoid the risks and seize the opportunities of using AI to support people with disability in the labour market.

Updated (2025): This work was followed by the next edition, including areas of vocational education and institutions, cognitive and sensory impairments.

Related

Key Findings and Statistics

Employment Gap Reality: In 2019, people with disabilities were 2.3 times more likely to be unemployed than people without disabilities across OECD countries, with their employment rate being 27 percentage points lower. This gap has remained stubbornly stable for the last decade, representing both equity concerns and significant waste of talent.

Technology Potential: The research identified 142 examples of AI-powered solutions that could support people with disability in the labour market, with over 75% of these solutions unable to exist without AI. This represents a fundamental shift from traditional single-purpose assistive technologies to versatile, adaptive solutions.

Areas and Categories

The report categorizes AI solutions into four distinct areas: Disability-Centred Solutions (60% of cases) focus on direct interventions including live captioning for deaf individuals, speech recognition for dysarthric speech, and AI-powered prosthetics. Environment Adaptation Solutions (25%) concentrate on making workplaces and content accessible, such as text simplification algorithms and accessible job-matching platforms. Meta-Level Accessibility Solutions (8%) work on improving processes that enhance accessibility, like workplace accommodation recommendation systems. Finally, New Work Opportunities (5%) create previously inaccessible employment opportunities, such as remote-operated logistics vehicles.

Repository

The attached repository reveals significant innovation diversity: 24% of solutions originate in academic settings, closely followed by small firms (23%), while big tech companies and startups represent 18% and 14% respectively. 87% are "first-intent" tools specifically designed to help people with disability, and solutions span multiple disabilities including vision (20%), hearing and motor (15% each), cognitive and speech (10% each). The repository demonstrates that over 75% of identified solutions would not exist without AI, meaning they require AI as an enabling technology. Technical approaches vary widely, featuring speech-to-text algorithms for live captioning, computer vision for environmental understanding and navigation, natural language processing for text simplification, brain-computer interfaces for motor control, and machine learning algorithms for personalized assistive technology recommendations. Notably, 45% of catalogued solutions remain in development phases, indicating the field's rapid evolution and significant growth potential.

Critical Challenges

Research & Development Barriers include lack of sustainable private funding beyond initial seed rounds, difficulty accessing relevant data and computing power, shortage of AI talent with accessibility expertise, and insufficient accessibility training among developers. Commercialization Obstacles encompass complex and lengthy reimbursement procedures, limited employer awareness of accessibility issues, discoverability challenges for emerging solutions, and difficulty establishing sustainable business models. Adoption Impediments feature the most cited barrier of lack of user engagement in solution development, along with infrastructure limitations and IT literacy gaps, and interoperability issues between new AI solutions and existing assistive technologies.

Expectations and limitations

While AI offers unprecedented opportunities—including cost reduction, greater personalization, and mainstream integration—the report cautions against techno-optimism. 45% of identified solutions are still in development, and experts warn that AI alone cannot address societal attitudes and stereotypes that perpetuate employment discrimination.

Recommendations

The report emphasizes that current policies are too fragmented and risk-focused. Key recommendations include:

  • Government-backed venture capital streams for accessibility-focused AI innovation

  • Simplified reimbursement mechanisms for AI-powered solutions

  • Mandatory accessibility clauses in public procurement

  • Enhanced data collection initiatives for inclusive datasets

  • Improved developer training on accessibility principles

The study concludes that realizing AI's potential for disability employment requires coordinated action across multiple stakeholders. Success depends not just on technological advancement, but on policy frameworks that encourage inclusive innovation, sustainable funding models, and meaningful engagement with disability communities throughout the development process.

This analysis demonstrates that while AI holds transformative promise for reducing the disability employment gap, achieving this potential demands strategic, sustained effort across research, policy, and implementation domains.

• • •

References

¹ OECD. "Using AI to support people with disability in the labour market: Opportunities and challenges." OECD Artificial Intelligence Papers. November 2023.

² OECD.AI Policy Observatory. "Labour markets." 2024.

³ OECD.AI Policy Observatory. "OECD Programme on AI in Work, Innovation, Productivity and Skills." 2024.

⁴ OECD. "OECD Employment Outlook 2023: Artificial intelligence and the labour market." 2023.

⁵ OECD. "Using AI in the workplace." Policy Report. March 2024.