Artificial intelligence (AI) is no longer reserved exclusively for programmers and tech companies. In the legal field, AI is increasingly being used for fact analysis, legal text interpretation, and even drafting legal advice. But instead of focusing solely on technological possibilities, I found it intriguing—being inspired by the book From Aristotle to Algorithm by Guido van der Knaap—to explore AI from a different perspective: that of the philosophical tradition. How does AI relate to fundamental concepts such as logic, knowledge, and causality?
Guido van der Knaap addresses these questions in an accessible way, placing AI within the context of six philosophical disciplines. Several themes stood out to me in relation to the legal profession.
Logic and Judiciary
Aristotle laid the foundation for legal argumentation with his formal logic. Legal reasoning revolves around carefully constructing arguments and drawing valid conclusions based on existing laws and precedents. Aristotle’s syllogisms—forms of reasoning where a conclusion logically follows from two premises—are still relevant in legal decision-making today.
AI can support this formal argumentation by analyzing patterns in legal texts, identifying contradictions, and reviewing contracts for legal consistency. Modern AI tools are increasingly capable of scanning legal documents and validating legal reasoning based on logical principles. This not only helps lawyers construct stronger legal arguments but also enables them to identify inconsistencies or weak points in reasoning in a timely manner.
Epistemology and Evidence
Epistemology is the philosophical study of knowledge: how do we acquire knowledge, what qualifies as reliable information, and under what conditions can something be accepted as truth? In the legal profession, epistemology plays a crucial role, as legal decision-making heavily relies on evidence. Lawyers and judges must determine which information can serve as compelling evidence, taking into account reliability, objectivity, and legal relevance.
AI, particularly Natural Language Processing (NLP), can make a valuable contribution to legal evidence-gathering by searching vast amounts of legal texts and quickly and accurately analyzing relevant case law and legislation. This not only speeds up the research phase but can also help uncover connections between rulings that might otherwise be overlooked. While AI still requires human interpretation, it is becoming increasingly proficient in assisting legal professionals with collecting and analyzing evidence.
The Limits of AI
Despite its impressive advancements, AI still faces philosophical and practical challenges. One of these is how AI understands cause-and-effect relationships. David Hume argued that we cannot directly observe causal relationships but only see events that frequently occur together. Deep learning models function in a similar way: they recognize patterns in data but do not necessarily understand the underlying causes.
This can be problematic in legal contexts, where the causal relationship between facts and outcomes is essential. For instance, an AI system might identify a correlation between certain case characteristics and the final verdict, but without a deep understanding of the legal principles underlying those decisions. Fortunately, AI models are improving in causal inference, enabling them not only to recognize correlations but also to better distinguish between causal relationships and mere coincidences.
A second fundamental challenge is the problem of induction. Hume pointed out that we tend to assume that patterns from the past will continue in the future, despite having no absolute certainty of this. AI models rely on induction: they learn from historical data and apply this knowledge to new cases. But what happens when circumstances change and existing patterns are no longer valid?
This can be problematic in legal and ethical contexts. An AI system trained on past case law may struggle to correctly interpret exceptional or novel situations. The solution lies in techniques such as meta-learning and transfer learning, which help AI become more adaptable and better suited to changing circumstances. This allows models to remain effective not only within their training data but also in new, unfamiliar contexts.
Conclusion
As the saying goes: “Without the past, there is no present; without the present, there is no future.” This also applies to the development of AI. AI builds upon past insights, operates in the present, and influences the future. As AI models continue to evolve at a rapid pace, their limitations are diminishing, and their reliability and relevance to the legal profession are increasing. If you, as a lawyer, want to understand AI’s impact on your field, be sure to take this book with you on your next vacation.
4o
