ORAL STATEMENT
Humanists International
Article 17 dialogue seminar at the European Parliament
“Health and well-being in the age of artificial intelligence:
communities tackling isolation and digital risks”
Speaker: Liam Whitton
Thank you Vice President Sberna.
I speak today on behalf of Humanists International, the global representative body for humanist, rationalist, secular, and laïque organisations in over 60 countries. This includes 28 member organizations in 21 EU countries.
It should come as no surprise that humanist organisations have taken a profound interest in artificial intelligence.
Broadly there are two reasons for this. The first is philosophical. If humanism, to slightly twist Protagoras’ words, takes man as the measure of all things, then AI opens up new windows on perennial questions about what it means to be human.
The second is more practical. Like many in this room, we have concerns about AI’s potential impacts on human beings, human wellbeing, and our individual capacities to lead happy and flourishing lives.
It is this latter concern which lies at the heart of the Luxembourg Declaration on Artificial Intelligence and Human Values, adopted by the global humanist movement just last year.
2.1
The Luxembourg Declaration states that the advent of AI cannot simply mean more people lining their pockets or new forms of surveillance affecting ordinary people. Properly shared, its benefit to individuals should include:
‘more time for leisure, fulfilling pursuits, learning, reflecting, and making richer connections with other human beings.’
Yet evidence shows that AI companion apps designed to address lack of human connection are proving, in the words of a 2026 study, ‘far inferior to a simple conversation with a complete stranger.’ Worse still, even these specially trained companions have repeatedly been linked to mental health deterioration and suicide attempts.
Under the AI Act and Digital Services Act, most chatbots are considered ‘limited risk’. But this is insufficient regulation for standalone digital companion apps marketed to lonely people. And it does not address hypothetical scenarios where AI companions are deployed in institutional settings.
Humanists in Europe have been providing emotional support and care to patients for over 80 years, and in that time, humanist pastoral carers and chaplains have learned that the art of listening – body language, tone, subtext, humour – includes, but extends well beyond, just finding the ‘right’ words.
I ask that MEPs recognise that the current legal framework has a gap. Rather than expand the availability of care, these technologies risk corroding the standards of human-centric care.
2.2
My second point relates to the Digital Omnibus on AI’s delaying the application of high-risk rules to December 2027 for standalone systems and August 2028 for regulated products.
These rules include a requirement that systems are designed around human oversight.
The idea that high-risk systems could go unregulated – including the role of AI in diagnostic, clinical, care, and triage decisions – for another 26 months is profoundly concerning.
The first principle of the Luxembourg Declaration asserts that
‘Decisions that deeply affect people’s lives must always remain in human hands.’
Far from keeping such decisions ‘always’ in human hands, new ‘grandfathering’ causes also mean that these rules may never apply to high-risk systems deployed before 2028.
By extending these deadlines (and who’s to say they won’t be extended again?), Europe risks allowing an entire technical and philosophical apparatus to take root in its hospitals. A new normal.
2.3
My third specific point relates to AI literacy and the risks posed by misinformation.
The Luxembourg Declaration warns explicitly that ‘human freedom depends on our ability to tell truth from falsehood’, and that legal frameworks must rise to meet the threat of AI-driven misinformation and ‘deliberate deception at scale’.
We only have to look back to the pandemic to see how consequential the spread of medical misinformation can be.
Unfortunately, the Omnibus weakens Article 4 of the AI Act by replacing an outcome-based duty to ‘ensure a sufficient level of AI literacy’ with a softer obligation to ‘take measures to support’ the development of AI literacy.
At the same time, enforcement of transparency and watermarking obligations for AI-generated content will also be delayed.
When conditions in Europe have rarely ever been more fertile for misinformation to spread, we risk being unprepared in the face of another health event like Covid-19.
The humanist novelist H.G. Wells once observed that as time went on,
‘Human history becomes more and more a race between education and catastrophe.’
We must prioritise education. Rather than weaken Article 4, we call on Members to work with the industry to set out a long-term vision for AI, health, and digital literacy. This could in turn help make compliance with Article 4 measurable and enforceable.
3.
The British humanist philosopher Bertrand Russell once famously said that ‘The good life is one inspired by love and guided by knowledge.’
AI will, however imperfectly, no doubt continue to aid in the expansion of human knowledge. It certainly has a role to play in the enrichment of human health and welfare.
But AI cannot supply love. It can simulate, but never embody, human empathy, human compassion, or human kinship.
These are human traits hardwired into us by our very survival as a species. Our evolution. We only exist today because, for all of our faults, human beings learned to love one another and care for each other.
‘Fairness by design’ must acknowledge the interconnectedness of health and happiness for every human being. This inevitably means greater caution where caution is the alternative to allowing harm.
‘Inspired by love, and guided by knowledge.’ I hope that something of this approach will resonate with MEPs in the days to come.
Thank you.
'Regulating AI for public health and well-being', Humanists International