Artificial Intelligence x Law x Strategy — April 2026

Three AI Models, One Case

Independent technical analysis by ChatGPT, Gemini and Claude on the same legal file — success probability, evidence strength and procedural risk assessment for a retrial petition.

How Did This Begin?

I started this process with very limited technical knowledge of social security law, armed only with documents and my own lived experience. I cannot write a legal petition like a specialist — but I realized that AI tools could help me understand which questions to ask, which legal provisions were relevant, and how to build a coherent argument.

I used ChatGPT for strategy, Gemini for technical analysis, and Claude for all 21 revision cycles of the petition — from v9 to v21. The result: a 14-page legal petition and a 42-page appendix document.

The core of the retrial petition: After the court judgment became final, the Social Security Institution (SGK) modified the claimant's service records. The accepted 6,330 days dropped to 3,580 — a loss of 2,750 days. This change is confirmed by the Institution's own official letter (Exhibit 4, 16.04.2026).

Lower Bound Estimate

Realistic Estimate

Upper Bound Estimate

Evidence Strength

The Dignity of Expertise

On the Legal Profession, the Necessity of Deep Specialisation, and the Meaning of Standing Firm — April 2026

There is a file before me. The written record of a struggle spanning years — countless petitions, shifting justifications, and judgments that consistently cut in the same direction. Reading this file, one cannot help asking: “Where did it go wrong?” The numbers are correct. The documents are in order. The premiums were paid — and then, after the judgment became final, the data was changed. The legal violation is numerically measurable and has been confirmed in an official letter.

So what is missing? The answer is simple but painful: A lack of specialisation. A guide who knows the law but who can also see precisely where that law intersects with this particular case, with this particular institution’s particular habits.

I. Compliance with Law and Compliance with Justice Are Not the Same Thing

What is taught in law school is the statute. But justice is the spirit of the statute — and these two concepts do not always coincide. A court can issue a decision based on the written rule; that decision can be both legally compliant and unjust at the same time.

A similar tension exists in this case. The Institution reaches the same conclusion each time through different justifications: the claimant cannot retire. First a gap-in-coverage claim, then a status reclassification, then a day-count issue, then an employer-share problem. The justification changes every time, but the outcome never does. This inconsistency may look legal; it is not just.

II. Being a Lawyer Is Not Enough — You Must Specialise

Every year thousands of lawyers graduate in Turkey. But the number of practitioners who are truly expert in one field — who know both the technical details and the institutional reflexes of that domain — remains insufficient.

Social security law is the most striking example of this. It appears straightforward on the surface: premiums, days, status categories. But once you enter it, you need to know the technical SGK regulations, the evolution of Supreme Court precedent, the transitional interpretations between the old Law No. 506 and the new Law No. 5510, and the binding hierarchy of regional appeals court decisions. Without all of this, it is impossible to manage this case correctly.

The root of the core problems in this case lies precisely here. Over the years, different lawyers, different perspectives, strategies that failed to build on one another. Each one a generalist; none of them a deep specialist in this domain. The result: technically strong facts that could not be translated into legal argument.

III. Justice Is Everyone’s Need — But Not Everyone Can Access It

Justice is not a luxury — it is a right. Yet in this country, and in much of the world, the cost of seeking one’s rights through the right lawyer, the right court, at the right time, far exceeds what many people can afford.

The person in this file paid premiums, registered coverage, and acted within the bounds of the law. Yet they are a decade into a battle. Because the system operates in a way that erodes the citizen’s patience and resources. The Institution has hundreds of lawyers, technical advisors, and legal databases for every procedure. The citizen is, more often than not, alone.

IV. Standing Firm — Refusing to Surrender to Despair

This case is not over. And it will not be — until either the right decision is reached or every legal avenue is exhausted. The retrial petition, then the Constitutional Court if necessary, then the European Court of Human Rights if necessary. This is not about refusing to give up — it is the clear and legitimate path of pursuing one’s rights.

Long battles wear people down. But there is a critical difference between despair and realism. Being realistic means saying: “This case has a 65% chance, there is risk, procedural interpretation is decisive.” Being despairing means saying: “It won’t work anyway, the system is broken, I give up.” The second position extinguishes the right itself.

Tools like the CharacterIX personality inventory remind us of this: resilience is not an innate trait but a value-based decision. If you know what your values are — in this case, being unjustly barred from retirement, paying premiums and not receiving the right they earned — then staying true to those values is a matter of character. And character reveals itself most powerfully under pressure.

V. Trust — In the Lawyer, the Law, the Future

Trust is the fundamental fuel of any legal system. The Institution told the same person different things over ten years. Courts reached different decisions on the same file. The regional appeals court issued two mandatory review orders — the trial court followed neither.

But this picture does not require abandoning trust. On the contrary: trust in the legal system is preserved by trying to correct it. The retrial petition is an expression of trust in the system’s own self-correction mechanism. I will share the outcome of this case — whatever it may be. Because this process is not merely one person’s retirement matter. It is a roadmap for others in similar situations who cannot find their voice. And the most concrete test yet of whether AI tools genuinely work in legal proceedings.

Conclusion: Growing and Standing

We must specialise. Being a lawyer is not enough; one must be someone who has worked in that field for years, who recognises the institution’s reflexes, who has tracked the evolution of case law in that domain. This is not merely an ethical obligation — it is a fundamental duty to the client.

It is also necessary to put on record: with limited technical knowledge, by asking the right questions and using three AI tools together, I prepared a 14-page petition and a 42-page appendix. Whatever the outcome, I have documented that this method works.

Every CharacterIX consultant sharing professional introduction pieces in their own area of expertise makes a powerful contribution to helping society understand the professions — and to guiding people toward the right kind of help.

April 2026 — Aykut YILDIRIM