Paul Murphy questions investigations into Grok and facial recognition

On 3 April 2026 Deputy Paul Murphy pressed Garda leadership over active investigations into child sexual abuse images linked to the XAI chatbot Grok and probed the use and governance of facial recognition technology. He asked whether investigations target the chatbot or its senior management, and raised concerns about bias and legal safeguards for biometric systems.

Investigations and accountability

Deputy Paul Murphy asked whether any of the 244 active investigations relating to images generated by the XAI chatbot Grok extend to the company or its senior management rather than only individual users. The exchange sought clarity on how the Gardaí will follow evidence that may point beyond users to platforms or companies, and on how transnational complexities could affect prosecutions.

Live facial recognition and cross-border questions

Murphy also asked whether Gardaí were informed after reports that UK police tested live facial recognition in Holyhead, noting the implications for a common travel area and effective border controls in practice. The Garda representative said Ireland currently lacks a legislative basis for live facial recognition and suggested such uses would be tightly constrained.

Retrospective biometric recognition: efficiency versus risk

The Gardaí described potential operational benefits from retrospective biometric recognition - for example speeding up manual reviews of thousands of hours of CCTV after incidents such as the November 2023 violence in Dublin. Murphy pressed on bias and accuracy concerns raised by the AI Advisory Council and asked what measures will be taken to reduce disproportionate error rates for certain groups.

Governance, evidential standards and human oversight

Garda witnesses stressed that any biometric tool would include human oversight and must meet evidential standards in court. Murphy focused on governance arrangements, safeguards and whether planned legislation gives sufficient protection against false positives, false negatives and amplified societal biases.

Consequences for policy and oversight

The exchange frames immediate policy questions for legislators and Garda governance: how to hold platforms or their management to account for AI-generated abuse imagery, how to regulate biometric systems, and how to ensure safeguards for affected communities while preserving investigatory capacity.