On Judgment, AI, and What We’re Not Seeing Yet - Issue #7
A recent set of comments from a senior judge on India’s Supreme Court has been circulating, framed around a familiar claim: artificial intelligence can assist the judicial process, but it cannot replace the act of judgment itself.
The reasoning is straightforward and, at face value, difficult to dispute. Judicial decisions require the weighing of nuance, the interpretation of context, and the ability to distinguish between cases that may appear similar on the surface but diverge in materially important ways.
The concern extends beyond capability into reliability. One of the more grounded critiques raised is that AI systems can produce outputs that resemble valid legal reasoning while relying on fabricated or non-existent case law. In other words, the system can sound authoritative without being anchored in reality, which introduces a specific kind of risk into any domain that depends on precedent and traceability.
There is also an implicit boundary being drawn. AI may support administrative efficiency, assist with drafting, or help organize large volumes of information, but the final act of judgment is understood to remain a human responsibility. The underlying assumption is that there is a clean separation between assistance and decision-making, and that this boundary can be maintained in practice.
That is where the conversation becomes less settled.
What is being debated publicly is whether AI can replace judgment. What is happening more quietly, across professional environments, is that AI is already participating in the formation of judgment. Not formally, not always explicitly, but in ways that are increasingly difficult to isolate.
When a legal associate uses AI to draft an argument, the structure of that argument is influenced before a human ever refines it. When a professional uses AI to summarize a body of material, the framing of what matters and what does not has already been shaped. When reasoning is tested against an AI system before being presented, the system is acting as a silent intermediary in the thinking process.
None of this constitutes replacement. It does, however, introduce a layer of influence that is rarely acknowledged as such.
The risk, then, is not simply that AI produces incorrect information. That problem, while serious, is at least visible. The more subtle risk is that AI produces reasoning that is coherent, persuasive, and incomplete in ways that are difficult to detect. When that reasoning is absorbed into human decision-making without clear attribution, responsibility becomes diffused.
The judge’s position assumes that judgment is a discrete act that occurs at the end of a process. In practice, judgment is often the accumulation of many smaller decisions, each shaped by the inputs available at the time. As those inputs begin to include AI-generated structures, summaries, and suggested lines of reasoning, the boundary between assistance and authorship becomes less distinct.
This is not an argument for or against the use of AI in judicial or professional settings. It is an observation about where the current framing may be insufficient.
If the question is limited to whether AI can replace judgment, the answer will likely remain no for the foreseeable future. If the question shifts to how AI is already influencing the conditions under which judgment is formed, a different set of considerations emerges—ones that are less about prohibition and more about visibility.
Before governance frameworks, oversight boards, or formal policies can be meaningfully applied, there is a more immediate need to understand where and how AI is already shaping decisions in practice. Without that clarity, attempts to regulate or contain its role will be built on an incomplete picture of its actual use.
The conversation, as it stands, is focused on the endpoint: the written judgment. The more consequential activity may be happening earlier, in the construction of thought that leads to it.
That is the layer that remains largely unexamined.
Rex C. Anderson
Desert Sage AI
AI Governance for Law Firms and Accounting Firms
__________________________________
Forwarded to you?
Subscribe for direct delivery of future issues.