Is AI the New DNA?

Is AI the New DNA ?

Chat-assisted Babak Taherzadeh

American Hero

Institutional Resistance, Finality, and the Next Evidentiary Reckoning

When DNA testing emerged in American criminal law, it was not welcomed as a neutral advance in truth-seeking. It was resisted—subtly at first, then openly—by courts, prosecutors, and institutions whose legitimacy depended on the finality of convictions. DNA threatened not only individual verdicts, but an epistemic assumption: that the legal system, once procedurally complete, had reached the truth.

Artificial intelligence now presents a similar disruption.

Courts today are increasingly engaged in efforts—formal and informal—to shield adjudicative processes from AI. Judicial ethics opinions warn against its use. Bar associations urge restrictions. Legislatures contemplate limits framed as “guardrails.” The stated concern is reliability, bias, or misuse. The unstated concern is far more fundamental: AI challenges the monopoly courts hold over post-hoc truth determination.

This moment closely resembles the early years of forensic DNA evidence.

I. DNA and the Myth of Finality

DNA testing did not invent wrongful convictions; it revealed them. Yet early judicial responses treated DNA claims as threats to finality rather than tools of accuracy. Courts questioned chain of custody, statistical interpretation, laboratory reliability, and—most tellingly—whether newly available testing justified reopening “settled” cases.

Finality, not truth, was the organizing principle.

Only after years of exonerations, public pressure, and undeniable empirical success did DNA evidence gain institutional legitimacy. Even then, its acceptance was narrowly cabined—often limited to biological evidence, strict procedural gateways, and high thresholds that excluded many meritorious claims.

The lesson is clear: transformative truth technologies are resisted not because they are weak, but because they are strong.

II. What AI Actually Does—and Why That Matters

Unlike DNA, AI is not a new type of evidence. It is an ability. It is analysis. Properly understood, large language models do not create facts; they organize, cross-reference, and evaluate existing ones. They expose inconsistencies, omissions, and structural defects that human reviewers—constrained by time, incentives, or institutional bias—often miss.

AI does not “believe” a conviction is final. It evaluates whether it is coherent, lawful, and free of corruption—and whethero fundamental rights were honored or denied and never corrected.

When provided complete, file-stamped records, AI systems flag contradictions between pleadings and orders, unexplained docket gaps, jurisdictional defects, and procedural impossibilities. In post-conviction contexts—where records are voluminous and stakes asymmetric—this capacity is uniquely powerful.

And uniquely threatening.

III. Why Courts Are Shielding Themselves from AI

The emerging resistance to AI in legal review mirrors early DNA skepticism, but with a critical difference: AI threatens systemic exposure, not merely case-by-case error.

DNA exonerated individuals.

AI reveals patterns.

It shows how records are altered, how filings disappear, how habeas petitions are denied without adjudication, and how procedural rules are selectively enforced. It highlights not just wrongful outcomes, but institutional habits.

This explains the urgency behind efforts to restrict AI’s role in legal contexts. The concern is not that AI will be wrong—it is that it will be consistently right in ways that undermine confidence in adjudication itself.

IV. AI as a Tool for the Justice-Denied

For individuals denied justice—particularly those barred procedurally from meaningful review—AI represents something unprecedented: a neutral, tireless auditor unconcerned with reputational risk or docket pressure.

AI does not defer to authority.

It does not rationalize error.

It does not confuse repetition with correctness.

Like DNA testing, AI offers those claiming wrongful conviction a way to ground their claims in demonstrable facts rather than credibility contests they are structurally doomed to lose.

This is why AI is increasingly framed as dangerous—not because it invents injustice, but because it documents it.

V. The Coming Reckoning

History suggests the pattern will repeat. Courts will resist. Gatekeeping doctrines will multiply. AI will be labeled unreliable, premature, or unethical. Meanwhile, its analytical power will quietly prove itself in the margins—used by journalists, scholars, litigants, and eventually, reluctantly, by courts themselves.

DNA did not replace trials.

AI will not replace judges.

But like DNA, AI may ultimately force the legal system to confront a truth it has long avoided: that procedural completion is not synonymous with justice, and finality is not a moral defense.

The question is not whether AI is the new DNA testing.

The question is how many years—and how many lives—it will take before the legal system admits it.