Testamentary capacity assessment in dementia using artificial intelligence: prospects and challenges

Area of Science:

• Forensic Psychiatry
• Cognitive Neuroscience
• Legal Capacity Assessment

Background:

• Testamentary capacity (TC) is crucial for valid wills, with rising importance due to aging populations and cognitive decline.
• Current TC assessments, based on the Banks v Goodfellow criteria, are subjective and do not solely rely on cognitive disorders.
• Situational complexity necessitates individualized assessment of a testator's circumstances.

Purpose of the Study:

• To propose a framework for an AI-driven decision support tool for testamentary capacity assessment.
• To address the interpretability challenges of AI in forensic psychiatry, particularly concerning GDPR.
• To integrate explainable AI (XAI) into capacity evaluation tools.

Main Methods:

• Development of a framework for an AI decision support tool.
• Integration of explainable AI (XAI) principles.
• Focus on statistical machine learning models adapted for TC assessment.

Main Results:

• A proposed framework for AI-assisted TC assessment.
• Emphasis on explainable AI (XAI) to ensure transparency and compliance.
• Potential for more objective and nuanced capacity evaluations.

Conclusions:

• AI, particularly XAI, offers a promising avenue for enhancing testamentary capacity assessments.
• The proposed framework aims to improve objectivity and address GDPR concerns in AI-driven legal capacity evaluations.
• Further research and development are needed to implement AI tools in forensic psychiatry for capacity assessment.