PhD- Computer Science (AI and Privacy) – Iceland

The University of Reykjavík invites applications for a fully funded PhD position in the Department of Computer Science. The position is part of the AI Antenna project, which brings together Nordic and European partners to advance fundamental research on artificial intelligence (AI) with strong societal relevance.

We are seeking a highly motivated PhD student interested in fundamental issues at the intersection of AI, privacy, and safety-critical systems.

The PhD project will investigate autonomous and intelligent systems (AIS) that preserve privacy, with a particular focus on de-identification and confidentiality in safety-critical domains. While healthcare will be the main area of application, the research is not limited to healthcare and aims to be generalised to domains such as public administration, law and finance.

Rather than focusing solely on existing privacy preservation techniques, the project will adopt a capability- and paradigm-independent perspective on privacy. The main objective is to identify privacy requirements and invariants that hold regardless of:

• Task (e.g., text processing, speech, vision, decision support),
• Data modality,
• AI method (e.g., machine learning, search, symbolic or hybrid systems).

The project is expected to make fundamental contributions to how privacy is defined, reasoned about, and applied in AIS, treating privacy as a property at the system and information flow level, not merely a training time constraint.

The research will be conducted in a robust research environment within RU and CADIA, RU’s interdisciplinary AI research centre, with opportunities for interdisciplinary collaboration and international engagement through the AI Antenna and CADIA networks.

The PhD is expected to be completed within three years, in accordance with the University of Reykjavík’s PhD regulations. 

Research topics may include (but are not limited to):

• Task- and modality-independent privacy principles for AIS
• Limits of de-identification and anonymisation in AIS
• Privacy risks arising from representation, inference, and temporal aggregation
• Adversarial and capability-based threat models for modern AIS
• Privacy guarantees beyond differential privacy, particularly for language and multimodal systems
• Design principles for privacy in safety-critical AIS, regardless of the learning paradigm

Requirements:

• Master’s degree in Computer Science or a closely related field
• Strong analytical and problem-solving skills.
• Solid background in the fundamentals of AI, machine learning, or related areas.
• Interest in privacy, security, or safety-critical systems
• Ability to work independently and as part of a research team.
• Excellent command of written and spoken English.
• Experience with natural language processing, speech, vision, or multimodal systems.
• Background in formal methods, information theory, or system-level reasoning.
• Previous experience in research or scientific publications.

Benefits:

• The position is fully funded for 36 months, including salary, and is exempt from tuition fees for the entire programme.

Organisation/company: University of Reykjavik.

Field of research: Computer Science .

Research profile: Early-stage researcher (R1).

Recognised researcher (R2)

Established researcher (R3)

Leading researcher (R4)

Country: Iceland.

Deadline for applications: 30 April 2026.

More information: Euraxess.