نوع مقاله : مقاله پژوهشی

نویسنده

• منصور انصاری

استادیار اندیشه سیاسی در اسلام، پژوهشکده امام خمینی و انقلاب اسلامی، تهران، ایران.

چکیده

با فراگیری استفاده از هوش مصنوعی، تحقیقات درباره نقش و اهمیت آن نه تنها در زندگی روزمره بلکه در همه حوزه‌های دانشگاهی شتاب گرفته است. در این مقاله از منظر علم سیاست به تأثیر هوش مصنوعی بر نظام‌های سیاسی می‌پردازیم. سؤال اساسی این است که هوش مصنوعی بیشتر به چه نوع نظام سیاسی کمک و یاری می‌رساند؟ در پاسخ به این پرسش تاکنون جواب‌های متفاوتی داده شده است. در اینجا فرضیه ما این است که هوش مصنوعی به احتمال بیشتر به تقویت و تحکمیم اقتدارگرایی منتهی می‌شود و به احتمال کمتری به تقویت دموکراسی منجر می‌شود مگر اینکه تمهیدات گسترده‌ای در نظر گرفته شود. نظریه هستی شناختی «لنگدون وینر» در خصوص ماهیت سیاسی مصنوعات بشری، اختراعات و ابداعات فنی بشری، است. بر مبنای نظریه وینر، پدیده‌های دست‌ساخته بشری خنثی نیستند و یا ذاتاً نوع خاصی از سیاست را تحمیل می‌کنند و یا اینکه دارای پیامدهای سیاسی مهمی هستند. به نظر ما، هوش مصنوعی با توجه به ماهیتش که تمرکزگراست به احتمال قوی به تقویت اقتدارگرایی منجر می‌شود و فقط با اتخاذ تمهیداتی در نوع هوش مصنوعی مورد استفاده و همچنین ایجاد واسطه‌هایی مانند حصارهای اخلاقی می‌توان آن را دموکراتیک‌تر کرد. تمهیداتی مانند توسعه الگوریتم‌های شفاف ساز، دموکراتیزه‌کردن زیرساخت‌های فناورانه، رقابت‌پذیر کردن پلتفرم‌ها، افزایش سواد داده‌ای شهروندان، و تدوین مقررات الزام‌آور اخلاقی. روش مقاله حاضر، روش تحلیلی استدلال ‌محور است. استدلال‌ها هرکدام از طرفین به دقت بررسی می‌شود و از آن‌ها در راستای دفاع از یک طرف استفاده خواهد شد. در نهایت، مقاله نتیجه می‌گیرد که هوش مصنوعی نه‌فقط یک ابزار فنی؛ بلکه نیرویی سیاسی است که بسته به چارچوب‌های حکمرانی، می‌تواند به تقویت یا تضعیف دموکراسی بینجامد. تعیین سرنوشت آن به تصمیمات نهادی، طراحی شفاف و مشارکت فعال شهروندان بستگی دارد.

کلیدواژه‌ها

• اقتدارگرایی
• حکمرانی الگوریتمی
• دموکراسی
• لنگدون وینر
• هوش مصنوعی
• هوش مصنوعی شفاف ساز

موضوعات

• علوم سیاسی

عنوان مقاله [English]

Artificial Intelligence, Authoritarianism, and Democracy

نویسنده [English]

• Mansour Ansari

Assistant Professor, Department of Political Thought, Imam Khomeini and Islamic Revolution Research Institute, Tehran, Iran.

چکیده [English]

Introduction
In recent decades, the emergence of Artificial Intelligence (AI) as one of the most transformative technologies of the 21st century has sparked extensive debates across the fields of politics, society, and philosophy. From general applications of artificial intelligence such as “ChatGBT” to surveillance and big data analysis systems, this technology has not only transformed human lifestyles but also challenged the foundations of political governance. In both public and professional discourse, conflicting views have emerged regarding the political implications of this technology: Is AI a tool to strengthen democracy, or is it a force for consolidating authoritarianism? And can an alternative, balanced model for its utilization be proposed? These questions have gained significance, particularly in light of theories such as Langdon Winner’s analysis of the politics of technology. In response to these concerns, the present article provides a multi-faceted examination of the relationship between Artificial Intelligence, democracy, and authoritarianism.
Research Objective
The main purpose of this research is to examine and analyze the arguments that consider AI as either serving authoritarianism or strengthening democracy. Adopting an analytical and theoretical approach, the author endeavors to demonstrate that, contrary to views that consider AI a “neutral tool,” this technology possesses inherent and structural consequences that potentially place it in the service of centralized and autocratic powers. Nevertheless, the research seeks to highlight possibilities for the democratic use of AI and propose solutions for mitigating its authoritarian ramifications. Consequently, the ultimate goal of the research is to contribute to a reconsideration of how novel technologies can be employed to reinforce the fundamental principles of democracy.
Research Methodology
This study is based on a theoretical-analytical method, relying on conceptual frameworks, philosophical arguments, and the analysis of political theories, rather than empirical data. Employing an argument-based approach, the research critically evaluates the pro and con perspectives regarding the relationship between AI and political systems. In this pursuit, the views of thinkers such as Langdon Winner, Shoshana Zuboff, Frank Pasquale, Ivan Krastev, Yuval Noah Harari, John Runciman, and others are assessed. The method of conceptual analysis is utilized to clarify key concepts such as “algorithmic governance,” “technological authoritarianism,” “human agency,” and “algorithmic transparency.” Classical political theories concerning power, transparency, and democracy are also employed to evaluate the technological implications.
Research Findings Report
The research findings indicate that AI, contrary to claims of neutrality, is structurally synergistic with authoritarian patterns. Five main areas supporting this have been identified:

Elimination of Human Agency and Political Subjectivity: As AI’s capacity for analyzing and predicting human behavior increases, conscious action and individual decision-making diminish. Zuboff explains with the concept of “cognitive alienation” how AI can bypass human awareness and will, distancing individuals from their own experience.
Mass Surveillance and Behavioral Control: Based on analyses by Zuboff and Bruce Schneier, mass surveillance has become a primary instrument of domination. By leveraging big data and AI, citizens are transformed into subjects of perpetual monitoring and control. This form of surveillance enables the exercise of power without the need for overt repression.
Obscurity and Lack of Algorithmic Transparency: Pasquale and other analysts have shown that many decision-making algorithms have a “black box” structure; meaning their decision-making process is neither visible nor assessable to the public, and sometimes not even to experts. This leads to the formation of authority without accountability.
Concentration of Data and Technological Power: As Krastev emphasizes, the accumulation and concentration of data in the hands of governments and large technology corporations have paved the way for technological dominance and software-based control over social and political processes. In China, this data concentration is intertwined with the engineering of public consent.
Globalization of Surveillance and Export of Authoritarianism: According to Harari’s perspective, countries like China can export their technology-based authoritarian model. In this model, regimes, by utilizing intelligent algorithms, generate consent and expand social control even without conscious citizen participation.

Alongside these arguments, the paper demonstrates that AI can also contribute to strengthening democracy under specific conditions. Four key areas highlight this potential:

Enhancing Transparency and Accountability: By designing Explainable AI (XAI) algorithms, the machine’s decision-making process can be illuminated, allowing for public assessment and oversight. Such processes can help rebuild public trust.
Strengthening Participation and Direct Democracy: AI-based platforms facilitate the collection of public opinion, feedback analysis, and the creation of effective communication channels between the government and citizens. Furthermore, policy feasibility can be assessed before implementation through sentiment and reaction analysis.
Increasing Awareness and Civic Education: Chatbots, law summarization tools, and personalized education via AI can enhance citizens’ civic awareness and political literacy. Especially in an era where data is abundant but processing it is difficult; AI plays the role of an awareness facilitator.
Combating Institutional Discrimination and Inequality: By employing machine learning and big data analysis, it becomes possible to identify patterns of discrimination in areas such as hiring, resource allocation, and administrative decisions. This process can help reduce institutional injustice.

Conclusion
The final results of the research show that AI is not merely a technical tool but a political structure that, in conjunction with power dynamics, can lead to either the consolidation of authoritarianism or the strengthening of democracy. Following Langdon Winner’s view, certain technologies inherently carry political imperatives. Therefore, the democratic use of AI necessitates institutional and technical interventions in five areas:

Developing transparent and explainable algorithms.
Democratizing access to technological resources and data.
Ensuring competition within the structure of large technology corporations.
Increasing citizens’ data and algorithmic literacy.
Formulating mandatory ethical and legal regulations.

Ultimately, the destiny of AI in relation to democracy or authoritarianism is determined not by its intrinsic nature but by its design, governance, and the level of public participation. In the words of Daniel Innerarity, AI can be both a threat to democracy and an opportunity for its reconstruction; everything depends on the choices we make.

کلیدواژه‌ها [English]

• Authoritarianism
• Algorithmic governance
• Democracy
• Langdon Winner
• Artificial Intelligence
• Explainable AI