Hosted by COFACE Families Europe and KMOP in Athens (Greece), and welcomed by a high-level panel including Dimitris Papastergiou (Minister of Digital Governance, Hellenic Republic), European and international community leaders and AI experts from 20+ countries met in Athens on 17th March to take stock of the state of play of artificial intelligence today, while also looking towards the future and its impact on families.
As artificial intelligence becomes increasingly embedded in daily life, its impact on families is growing, shaping how they work, learn, care, communicate, and stay safe online. The European expert meeting convened researchers, policymakers, tech experts and civil society to discuss the impact of artificial intelligence on family life and inclusion, with a focus on rights, equity and empowerment in the digital age.
Adopting a multigenerational, child rights and disability-inclusive approach, the expert meeting provided a platform for European family organisations and stakeholders from the COFACE Families Europe network to exchange insights on families’ experiences across different areas of daily life, including the home, the labour market, schools and the online environment, while examining both the opportunities and potential risks it presents for families. Speakers from local to international level provided both positive examples (e.g. the use of AI to support family reunification as presented by the Turkish Red Crescent) and negative examples (e.g. as highlighted by the 5Rights foundation on high risks for children online). Fresh evidence was presented by the OECD and the Institute for Structural Research (Poland) on the risks and opportunities for AI-exposed workers, by the European Commission on Generative AI in secondary schools, as well as research highlighting how the integration of Artificial Intelligence (AI) in social services is transforming how these services are delivered.
Some key points raised during the discussions include the following:
- Artificial intelligence tends to be used as a general term, lumping all AI systems into having the same risks. It is essential for the COFACE community to be clearer about what we are talking about so that we can name the risks/opportunities for each type of AI: algorithms, generative AI, facial recognition and more. See here a glossary of basic AI terms. A delegation of ambassadors from Better Internet for Kids Youth were present to reflect together with the participants on the impact of AI on youth, highlighting that AI literacy is essential moving forward.
- Discussions about social media bans and AI systems tend to be separate, but they are intrinsically linked. Social media platforms use addictive algorithms (a form of AI) to keep children and youth active on their platforms. This is not sustainable: the tech industry have a duty of care and safeguarding responsibilities to ensure that children and youth using their products have a safe experience. There were many strong calls for the tech industry to conduct proper age-verification on their products/platforms and implement the current societal standards in place in Europe for children below 13 and between 13-15 years.
- The backdrop to all the discussions was the EU regulatory framework, with an in-depth assessment through the COFACE families lens, focusing on different regulations and a specific focus on the Artificial Intelligence Act (Act).
- Many participants had never heard of the AI Act, with recommendations to ensure the AI Act reaches the everyday conversations of families at home (like the GDPR).The AI Act aims to protect health, safety, and fundamental rights, which include non-discrimination and equality. It contains safeguards for protection of minors online such as the prohibition of AI systems that exploit vulnerabilities based on age (Article 5) and classification of high-risk AI systems (Article 6). The Act also includes safeguards in relation to financial inclusion with two key financial-sector applications covering personal loans and mortgages, and risk assessment and pricing of life- and health insurance. The Act is disability-inclusive, with mandatory accessibility requirements built into the law by design, including strong requirements for high-risk AI systems to be usable by persons with disabilities. Furthermore, persons with disabilities are explicitly recognised as a group at higher risk of harm from AI systems.
- However, there are still many regulatory gaps affecting families. There are many regulatory gaps such as the lack for a clear prohibition of “nudifying tools” in the AI Act. Given the scale, permanence and severity of harm linked to AI-enabled nudes and child sexual abuse imagery, such a prohibition is both proportionate and necessary. As regards families in migration, the AI Act leaves broad discretion for Member States to deploy AI in migration decisions. Finally, existing EU sectoral financial laws were approved before AI systems even existed or were widely used, and there are no EU harmonised redress and liability rules tailored to AI systems.
- A new EU proposal coming from the European Commission (referred to as the “Digital Omnibus”) is putting human rights, including children’s rights, at risk by proposing to remove transparency safeguards for high-risk AI Systems, undermining the AI literacy obligation, and delaying implementation of the AI Act (including on watermarking deepfakes). It is not possible to make the protection of minors in the digital environment a European political priority and talk about social media bans, while failing to regulate AI and delaying the implementation of strong safeguards.
- There was a general call to action to strengthen the AI Act, as a way to strengthen families. Artificial intelligence is not only transforming economies and institutions. It is quietly transforming family life. How can artificial intelligence strengthen families, rather than deepen inequalities or vulnerabilities? A first step to ensure ethical, inclusive and empowering AI systems for families is to enforce strong EU regulatory frameworks: starting by a strengthening, not weakening, of the AI Act for the sake of innovation. The current proposed ‘simplification’ of the AI Act is reducing the accountability of companies and certainly not simplifying the rules to better protect European families.
- Calls for Family-Centred AI: There were many calls from participants to the tech industry to adopt a family-centred approach in AI design, deployment and governance, prioritising human dignity and relationships while supporting families’ capacity to care, provide, and belong.
See further information about the meeting here.
