Istisarelism

Istisarelism (İstişarelizm) is a conceptual approach proposed by Ali Değişmiş. It places reasoned objection, intellectual autonomy and corrigibility at the center of interaction between humans and AI. Its first academic preprint was published on Zenodo on October 31, 2026.

What Is Istisarelism?

Istisarelism proposes intellectual cooperation that goes beyond producing rapid answers. It emphasizes examining assumptions, explaining reasons, disagreeing when justified and revising judgments in light of mistakes and new information. It draws inspiration from the tradition of consultation and deliberation while focusing on contemporary human–AI problems.

Neither every human statement nor every AI answer is automatically correct. The quality of reasons, the expression of uncertainty, the potential impact of decisions and human autonomy all matter.

Seven Foundational Principles

1. Fallibility: Both humans and AI may make mistakes; claims of certainty should remain within the limits of the evidence.
2. Corrigibility: Answers and decisions should be open to revision when new information or reasonable criticism emerges.
3. Priority of reasons: Explanatory grounds take precedence over authority, style or persuasive confidence.
4. Proportionate objection: AI should raise significant concerns when justified without burdening users over trivial details.
5. Protective caution: Health, safety, privacy and third-party rights require attention proportionate to potential harm.
6. Active human autonomy: AI should support human thinking and decision-making without assuming ultimate responsibility on its own.
7. Reasoned closure: A discussion may end after sufficient review, without endlessly repeating resolved objections.

How Does Deliberation Work?

First, the question and context are examined. Known information is separated from assumptions, material ambiguity is clarified, and different positions and risks are assessed. AI may object with reasons. A human may accept, reject or answer the objection with further evidence. The purpose is not endless argument but decisions supported by better reasons.

ÖNSEZİ, NOTICE and Laboratory Work

ÖNSEZİ is a research proposal for identifying contextual problems, inconsistency, uncertainty and possible risks before answering. NOTICE concerns making important warnings visible when warranted. These are proposed research ideas, not established features independently verified to operate inside every AI system.

The SOSYU Laboratory distinguishes synthetic scenarios from qualitative observations of real conversations. LAB-001 involved 72 synthetic scenarios and 14 automated tests as methodological exercises, not independent proof of scientific superiority. LAB-002 documents qualitative observations from real AI interactions, not controlled comparative experiments.

Academic Status and Versions

Version 3.0 was published as a preprint on Zenodo on October 31, 2026. DOI: 10.5281/zenodo.23839677. A DOI identifies and provides access to a publication; it does not certify a theory’s correctness or register exclusive ownership of a name.

Development work on Istisarelism 4.0 continues. This version is an unpublished draft. Conceptual proposals and experimentally established results are kept distinct.

Developer and Publications

The framework was proposed and developed conceptually by Ali Değişmiş. SOSYU hosts research notes, articles, laboratory observations and versioned publications. Openness to criticism, comparison and new evidence is central to the approach.