On 28 October 2025, the European Disability Forum (EDF) hosted an online event to present the updated version of its AI Act Implementation Toolkit. The session came at a pivotal time, as the European Union’s Artificial Intelligence Act moves from its legislative phase into implementation. For autism and disability advocates, the coming months will be crucial in ensuring that the Act’s promises of fairness, accessibility, and transparency are realised in practice.
The AI Act formally entered into force on 1 August 2024, establishing the world’s first comprehensive framework to regulate artificial intelligence according to levels of risk. The Regulation is to be rolled out in stages, with the first provisions applicable since February 2025. These include bans on certain “forbidden” AI systems, such as those involving harmful manipulation, social scoring, or exploitation of vulnerabilities. By 2 August 2025, Member States must have designated their national authorities responsible for implementation, including market surveillance agencies and equality or human rights bodies under Article 77. A list of the single points of contact of each member state is available on the EU Commission’s website: Market Surveillance Authorities under the AI Act. The Act will become fully applicable in August 2026 with some exceptions, including regarding the rules for high-risk AI systems embedded into regulated products, which now have an extended transition period until 2 August 2027.
Strengthening National engagement through the updated Toolkit
These milestones set the timeline for organisations across Europe to influence national processes and ensure that the disability perspective is reflected in how the law is applied. The updated EDF Toolkit (link below) is designed precisely for this purpose: to equip civil society with the information and resources needed to engage effectively at national level. The updated version builds on the 2024 edition with expanded sections on governance, monitoring, and advocacy. It offers practical explanations of how AI systems are categorised by risk within the legislation, and how disability organisations can take part in consultations, contribute expertise, and report non-compliance. This new version is in response to the guidelines on the legislation published by the European Commission.
One important update is the focus on the responsibilities of national authorities. The toolkit explains the role of market surveillance authorities, which will oversee compliance and carry out investigations, and equality bodies, which can provide human rights expertise and request documentation when fundamental rights are at stake. Another addition covers how organisations can lodge complaints when AI systems fail to meet accessibility or bias-testing requirements.
A National Model: CERMI’s Comprehensive AI Governance Strategy
The webinar highlighted an example of structured national advocacy through the presentation from CERMI, Spain’s national umbrella organisation of persons with disabilities, which is developing a comprehensive AI governance strategy for the disability sector. CERMI outlined a multi-year framework that brings together governance, risk management, regulatory compliance, and AI literacy, with the aim of ensuring that all AI systems used within the organisation align with both legal requirements and ethical standards by 2026. The approach includes building internal capacity to understand the impacts of AI, creating an inventory of AI systems in use, piloting responsible AI projects that support rights-based advocacy, and positioning the organisation as a leading actor in shaping inclusive digital policies at national level.
Implementation of the Act for Autistic people
For autistic people, the stakes are particularly high. AI systems increasingly shape decisions about education, employment, health care and social services—areas where bias or misunderstanding can have profound effects. Emotion-recognition and behavioural analytics tools, for example, may misinterpret autistic communication or social interaction, leading to misjudgements or exclusion. The AI Act’s safeguards on accessibility, transparency, and bias testing are therefore essential to protect against discrimination and to promote inclusive design. The toolkit provides autism and disability organisations with a concrete means to monitor these systems and advocate for fairer AI governance.
The changing EU Digital Policy Landscape
The webinar also addressed the challenging policy environment in which the AI Act implementation is taking place. Blue Tiyavorabun, representing European Digital Rights (EDRi), outlined the growing political momentum within the EU for what is being described as “competitiveness and simplification”, a euphemism that often translates into deregulation efforts. The European Commission’s forthcoming “Digital Omnibus” package, expected in November 2025, is part of a wider trend of rolling back or delaying human rights protections in digital law. As noted in a recent analysis by Tech Policy Press (“EU Set the Global Standard on Privacy and AI – Now It’s Pulling Back”), Europe’s reputation as a global leader in rights-based digital regulation is at risk of erosion under mounting industry pressure and political shifts. These developments could affect not only the AI Act, but also other landmark frameworks such as the GDPR, ePrivacy Regulation, and wider data governance initiatives.
Against this backdrop, EDF called for sustained vigilance and cross-sector cooperation. Disability organisations were encouraged to establish early contact with their national AI Act enforcement authorities and equality bodies, to participate in national consultations, and to collaborate with consumer and digital rights groups.
Advocacy in this next phase
For the autistic community, this phase of the AI Act presents both challenges and opportunities. The law sets important precedents for accessibility and fairness, but its impact will depend on how robustly these standards are enforced. Autism-Europe and its members can play a vital role by contributing expertise, monitoring national developments, and raising awareness of rights such as the “right to explanation” for those affected by high-risk AI decisions.
This article continues Autism-Europe’s ongoing coverage of AI and disability policy. Readers can revisit our earlier feature, Advocating for Autism in AI Regulation: Navigating Europe’s New AI Act (October 2024), available here, which introduced the AI Act and the first edition of the EDF toolkit.
As the implementation phase unfolds, active engagement from the disability community will be essential in defining how AI serves or harms people across the EU. With the updated EDF toolkit in hand and strong advocacy at national and European levels, autism and disability organisations can help ensure that the future of AI in Europe is inclusive, transparent, and grounded in human rights.