نوع مقاله : مقاله پژوهشی

نویسنده

• سید مهدی مجیدی نظامی

استادیار، گروه سازماندهی اطلاعات، پژوهشکده مدیریت اطلاعات و مدارک اسلامی، پژوهشگاه علوم و فرهنگ اسلامی، قم، ایران

چکیده

در بحبوحه رقابت جهانی بر سر تنظیم‌گری هوش مصنوعی، این پژوهش مدل «حکمرانی درونی» را به‌طور انتقادی تحلیل کرده است. مسئله اصلی، عبور از دوگانه «تنظیم‌گری دولتی» در برابر «خودتنظیمی» و واکاوی سازوکارهای شرکت‌های بزرگ فناوری برای تبدیل استانداردهای فنی به ابزار قدرت سیاسی و اقتصادی است. این پژوهش کیفی
با بهره‌گیری از روش «تحلیل گفتمان انتقادی» و در چارچوب اقتصاد سیاسی ارتباطات انجام شده است. داده‌ها
شامل اسناد رسمی، فنی و سیاستی سه شرکت پیش‌روی آمریکایی (Open AI Google DeepMind, Anthropic, ) به‌عنوان معماران اصلی گفتمان «ایمنی» و مدل حکمرانی مبتنی بر سرمایه‌گذاری خطرپذیر است. حکمرانی درونی از طریق معماری سه‌لایه فنی، سیاستی و سازمانی اعمال می‌شود و سه کارکرد اصلی دارد: ۱) ایجاد موانع ورود از طریق افزایش هزینه‌های انطباق، ۲) «تسخیر نرم مقررات» برای پیش‌دستی بر قانون‌گذاری دولتی، و ۳) تثبیت «هژمونی الگوریتمی» که ارزش‌های اکوسیستم فناوری آمریکا را جهانی‌سازی می‌کند. حکمرانی درونی یک پروژه ژئوپلیتیک برای حفظ انحصار فناورانه است. نوآوری پژوهش، طراحی «الگوی سوم حکمرانی» از طریق ارائه استاندارد پیشنهادی «الگوریتم پارسا» به‌عنوان بدیلی بومی در برابر استانداردهای سیلیکون‌ولی است. این استاندارد چارچوبی فنی برای ممیزی و سنجش انطباق هوش مصنوعی با ارزش‌های بنیادین ایران و گذار از «مصرف‌کننده منفعل» به «حکمران فناورانه فعال» فراهم می‌کند.

کلیدواژه‌ها

• حکمرانی
• هوش مصنوعی
• استاندارد الگوریتم پارسا
• حاکمیت فناورانه
• اقتصاد سیاسی فناوری
• هژمونی الگوریتمی

عنوان مقاله [English]

From Algorithmic Hegemony to Technological Sovereignty: A Critique of the Political Economy of the Internal Governance of Artificial Intelligence and the Design of the Proposed "Parsa Algorithm" Standard

نویسنده [English]

• Sayyed Mahdi Majidi Nezami

Assistant Professor, Department of Information Organization, Research Center for Islamic Information and Document Management, Islamic Sciences and Culture Academy, Qom, Iran

چکیده [English]

Amidst the global competition over "artificial intelligence regulation," the present article critically examines a model it terms "internal governance"; a model in which big tech companies are not merely subject to external rules, but are themselves the designers of norm-setting and standard-setting mechanisms in the field of artificial intelligence. The central problem of the research is to transcend the conventional dichotomy of "state regulation" versus "self-regulation"; because this dichotomy alone is incapable of demonstrating how ostensibly neutral technical standards can be transformed into instruments of political and economic power. The article argues that leading companies, by formulating discourses such as "safety," and relying on internal technical, policy, and organizational mechanisms, turn technical standards into mechanisms of exerting influence; in such a way that its consequence can be the formation of "algorithmic hegemony" and the consolidation of technological monopoly. From this perspective, the issue is not merely the regulation of a technology, but rather the power relations at the level of the political economy of technology and its geopolitical consequences: who writes the rules of the artificial intelligence game, how they globalize it, and place others in a position of subordination. The primary objective of the research is the critical analysis of "internal governance" in the domain of artificial intelligence and the exploration of the mechanisms through which big tech companies transform technical standards into instruments of political-economic power. Alongside this explanatory objective, the article also seeks to present an alternative: designing a "third model of governance" through proposing the "Parsa Algorithm" standard as a localized option against the dominant standards of Silicon Valley; an option that can provide a technical framework for auditing and assessing the compliance of artificial intelligence with the "fundamental values of Iran" and strengthen the possibility of transitioning from the status of a "passive consumer" to an "active technological governor." The current research seeks to answer the problem of: how does 
the "internal governance" of large artificial intelligence companies transcend the state regulation/self-regulation dichotomy through transforming technical standards into mechanisms of influence, and lead to the creation of political-economic power and the consolidation of algorithmic hegemony? And based on this, how can the "third model of governance," relying on the proposed "Parsa Algorithm" standard, create a framework for auditing and value-orienting artificial intelligence in Iran as a localized alternative? This is a qualitative research conducted using the "critical discourse analysis" method and formulated within the framework of the "political economy of communications." The research data include the official, technical, and policy documents of three leading American companies—Open AI, Google DeepMind, and Anthropic—which are introduced in the article as the principal architects of the "safety" discourse as well as the governance model based on venture capital. The methodological focus of the research is on extracting the discursive logic and institutional mechanisms that, through documents and procedures, lead to standardization and rulemaking affecting the public policymaking environment. The findings of this research indicate that the internal governance model of big artificial intelligence companies is implemented through a three-layered technical, policy, and organizational architecture. This mechanism allows companies to restrict the entry of new actors into the technological ecosystem by creating structural barriers and increasing compliance costs. Internal governance, through the "soft capture of regulations," preemptively shapes the state legislative space and globalizes its desired standards in the form of the safety discourse and the values of the American technological ecosystem, so that formal regulations are often compelled to comply with the framework of leading companies. Consequently, algorithmic hegemony is formed; meaning that the values and criteria determined by American companies are accepted as a global model and indicator, and other countries and technological actors are driven to follow them. This mechanism, in addition to economic and policy dimensions, is also subject to geopolitical logic and operates as a project to preserve Western technological monopoly. Accordingly, the innovation of the research lies in introducing the "third model of governance" and the localized "Parsa Algorithm" standard, which provides a way to assess and audit artificial intelligence based on the fundamental values of Iran, and to transition from a space of passive consumerism to active technological sovereignty. The results suggest that focusing on the dichotomy of "state regulation" versus "self-regulation" is insufficient for understanding the reality of artificial intelligence governance; because big companies transform technical standards into instruments of exercising power through the path of "internal governance" and with a three-layered technical-policy-organizational architecture: both by raising compliance costs, by preempting the shaping of regulations, and by consolidating the algorithmic hegemony that globalizes the values of the American technological ecosystem. Against this situation, the proposal of the "Parsa Algorithm" as a localized standard opens the path for designing the "third model of governance" so that the possibility of technical auditing of artificial intelligence based on the fundamental values of Iran is provided, and the country moves from the role of a "passive consumer" toward "active technological sovereignty."

کلیدواژه‌ها [English]

• Governance
• Artificial intelligence
• Parsa algorithm standard
• Technological sovereignty
• Political economy of technology
• Algorithmic hegemony