
Artificial Intelligence in Social Security
Conference highlights the synergy between AI and the European Pillar of Social Rights.
HS – 03/2024
The application of artificial intelligence (AI)
in the field of social security and in particular in the assessment of work
capacity was the focus of the conference entitled "Shaping
Tomorrow: AI in Social Security and Work Capacity Assessment". The event
took place on 6 March and was organised by the European Union of Medicine in
Assurance and Social Security (EUMASS) in collaboration with the Belgian
Ministry of Social Affairs and supported by the Belgian Council Presidency of
the European Union (EU).
In principle, the organisers of the symposium
sought to link the topic of AI with the European
Pillar of Social Rights (EPSR). The aim
was to showcase the synergy between AI and the 20 principles of the EPSR and
to contribute to an interdisciplinary dialogue to explore the opportunities and
risks of AI in this area.
Challenges and potential
The four panel discussions focussed on different
topics. After an introductory panel with the Belgian Minister of Social Affairs and Public Health, Dr
Frank Vandenbroucke, the second panel discussion focussed on the topic of (health‑) data and AI in social
security. The third panel discussion focussed on the challenges of using AI in social security and in
the assessment of work capacity, while the fourth finally focussed on the benefits and future potential of AI for
work capacity assessment.
Data plays a key role
There was agreement on the demand for access to
representative data. AI applications are neither neutral nor objective by nature.
They must be trained using large databases relevant to the problem. The added
value of AI applications lies in the constantly improving analysis of large
amounts of data. This would require comprehensive, up-to-date and, above all,
representative data. Otherwise, there is a risk that AI applications will
reinforce existing prejudiced decisions in the analogue world. In this context,
the importance of the Directive
for a European Health Data Space (EHDS) was also emphasised. In order to increase the potential of AI
applications, in addition to access to existing social data, a systematic,
standardised recording of the results of, for example, medical or occupational
rehabilitation measures is also required.
AI in the healthcare sector
The conference generally came to the conclusion
that AI should be seen as a transformative tool in the healthcare sector that
could improve the accuracy of work capacity assessments and counteract existing
inequalities in access to medical and occupational rehabilitation. Due to the
susceptibility of AI-based applications to incorrect decisions due to a
potentially incomplete or distorted data basis, all panellists agreed that the
use of AI-based tools should be accompanied by an open mind, but should always
be critically questioned.
Social impact of the AI Act
A number of aspects recurrently came up during the conference, including the EU AI Act and the topic of transparency. Although there are provisions on
this in the AI Act, it is not always clear how AI-based applications reach
decisions in the first place. The demand for transparency therefore runs the
risk of amounting to nothing. Transparency and accountability are essential for
public trust and thus the acceptance of AI applications. Following the adoption
of the AI Act, the social impact should therefore be closely monitored.
Furthermore, social partnerships and dialogue should not be lost sight of in
order to discuss the use and expectations of AI.