adopted by the Board of Directors of the European Disability Forum on the 1st of April 2023
Highlighting that Artificial intelligence (AI) is an emerging technology that has made its way into various areas of our lives, such as personalised recommendations online, virtual personal assistants, chatbots, speech-to-text technologies and computer vision applications, and AI based applications are also used by public and private services, including in the areas of social policies, healthcare, insurance, banking, education and employment.
Recalling that emerging technologies such as AI can further support the social participation and inclusion of persons with disabilities if they are developed with a Design for All approach, and used in a way that respects the right to non-discrimination, equality and privacy.
Concerned that persons with disabilities are at increased risk of becoming victims of fraud or misinformation through misuse of generative AI[1], and limitation of their self-determination.
Stressing that AI solutions that are not designed taking human diversity into consideration can increase the exclusion and discrimination against already disadvantaged groups, such as persons with disabilities, but also women, racialised persons and ethnic minorities, older persons and other discriminated people. Such discrimination can occur without the person noticing it.
Recalling that the United Nations Convention on the Rights of Persons with Disabilities (CRPD) has been the first international human rights instrument that recognises the need to promote equal access to technology as a fundamental right for persons with disabilities (article 9) and to ensure equality and non-discrimination in all areas of life (article 5).
Underlining that the United Nations Special Rapporteur on the Rights of Persons with Disabilities recognised that the negative impacts of AI on the rights of persons with disabilities must be openly acknowledged and addressed through cooperation between states, companies, national human rights institutions, civil society, and organisations for persons with disabilities.[2]
Stressing that a lack of accessibility in the design and use of AI can endanger the lives and well-being of persons with disabilities.
Noting that the EU Strategy for the rights of persons with disabilities 2021-2030 recognises that accelerated digital transformation and the green transition offer opportunities, using information and communication technology (ICT), artificial intelligence and robotics to design on-site and remote services tailored to the needs of persons with disabilities; but that the effective use of these technologies requires the removal of accessibility barriers for persons with disabilities and investing in their digital skills.
Welcoming the decision of the European Commission to propose an EU law – the AI Act – to regulate the use of AI in three risk categories, while concerned that the proposal does not sufficiently include and protect the rights of persons with disabilities.
The EDF Board of Directors calls on the European Union institutions to:
- Live up to its obligations under the CRPD and ensure that all persons with disabilities can benefit from accessible, affordable, and available AI-technologies that support their socio-economic participation and independent living, and that AI systems put in place in the EU and worldwide respect persons with disabilities’ rights to non-discrimination, equality, self-determination, and privacy.
- Adopt the proposed EU Regulation on AI ensuring that AI systems, regardless of their level of risk, are subject to mandatory accessibility requirements. This includes AI-related information and user guides, which will need to be in line with existing EU legislation on accessibility, in particular the European Accessibility Act.
- Recognise the disproportionate risk to the rights of persons with disabilities in the areas identified as high risk in the proposed AI Act, such as biometric identification, access to education, employment, private and public services, law enforcement, migration and border control, and the administration of justice.
- Enlarge the list of prohibited uses of AI systems to cover, in particular:
- Biometric identification and categorisation of natural persons.
- AI systems that determine the possibilities for individuals to gain access to education and employment.
- Access to and use of essential services and benefits, both private and public, including provision of healthcare and emergency services (Except for prioritizing emergency response services – see page 9 in the EDF position paper).
- Use of AI by law enforcement (Except for detecting deepfakes – manipulated video, audio or image that looks real but is digitally created – and to evaluate the reliability of evidence – see page 9, in the EDF position paper)
- Use in the management of migration, asylum and border control. (Except for evaluating the authenticity of traveling documents and assisting authorities in examining applications and associated complaints – see page 9, in the EDF position paper)
- Ensure that the Regulation protects the privacy and data protection of all persons with disabilities, including persons with intellectual and psychosocial disabilities and persons subject to substitute decisions, such as guardianship, when their data are processed by AI systems.
- Put in place effective and accessible measures so that individuals, including persons with disabilities, are informed when their data is being collected and have the opportunity to inquire about and object to the processing of such data.
- Prohibit public and private entities from using AI to recognise emotions, except for certain well-defined research purposes subject to strict privacy safeguards, including informed consent and the ability of research subjects to object. AI-based emotion recognition assistive technologies should also comply with such safeguards.
- Ensure the protection of individuals’ fundamental rights in relation to the use of AI, including measures for reporting problems, making complaints to competent authorities (including collective complaints and complaints made by civil society actors on behalf of individuals) and seeking remedies in the event of abuse, and ensure that these measures are accessible to persons with disabilities.
- Promote the development of AI that involves the meaningful participation of experts with disabilities, accessibility professionals and other rights holders through financial and other incentives (such as EU and national funding for projects led by disability organisations or directly involving accessibility professionals).
- Adopt measures to support the development of AI for human and societal benefit, taking into account the needs of diverse communities, including persons with disabilities.
- In accordance with Article 4(3) of the CRPD, to closely consult and actively involve persons with disabilities, through their representative organisations, in the development, implementation and monitoring of European and national AI policies, including in relation to the EU Coordinated Plan on AI and national AI strategies.