فصلنامه نوآوری حقوقی

فصلنامه نوآوری حقوقی

بزه‌انگاری تهدیدات نوظهور ناشی از هوش مصنوعی: تحلیل چالش‌ها و راهکارهای تقنینی

نوع مقاله : مقاله پژوهشی

نویسندگان
1 دانشیار گروه حقوق کیفری و جرم‌شناسی، دانشکدۀ حقوق و علوم انسانی، دانشگاه کاشان، کاشان، ایران
2 کارشناس ارشد حقوق کیفری و جرم‌شناسی، دانشکدۀ حقوق و علوم انسانی، دانشگاه کاشان، کاشان، ایران (نویسندۀ مسئول)
3 کارشناسی ارشد حقوق کیفری و جرم‌شناسی، دانشکدۀ حقوق و علوم انسانی، دانشگاه کاشان، کاشان، ایران
چکیده
گسترش سریع فناوری‌های مبتنی بر هوش مصنوعی، چالش‌های تازه‌ای را در حوزه‌های اجتماعی، امنیتی و اقتصادی پدید آورده است. این تحولات، پرسش‌هایی را دربارۀ میزان کارآمدی چارچوب‌های موجود حقوق کیفری ایران در مواجهه با این پدیده‌ها مطرح می‌کند. مسئلۀ اصلی این پژوهش، بررسی چالش‌های بزه‌انگاری تهدیدات ناشی از هوش مصنوعی و تحلیل خلأهای تقنینی موجود در نظام حقوقی ایران است. این مطالعه با رویکردی نظری ـ تحلیلی و بر پایۀ منابع معتبر داخلی و بین‌المللی، به واکاوی تهدیدات علیه اشخاص (مانند جعل عمیق و فیشینگ)، تهدیدات ناشی از فناوری‌های خودران و خودمختار (مانند خودروهای خودران و سلاح‌های خودمختار مرگبار)، تهدیدات علیه امنیت (حملات سایبری به زیرساخت‌های حیاتی) و تهدیدات علیه اموال (باج‌افزار و استخراج غیرقانونی ارزهای دیجیتال) می‌پردازد. یافته‌ها نشان می‌دهد قوانین موجود در زمینۀ جرم‌انگاری رفتارهای فناورانۀ ناشی از هوش مصنوعی، پراکنده و ناکافی بوده و توان پاسخ‌گویی به پیچیدگی‌های تهدیدات نوظهور را ندارند. نوآوری پژوهش در تحلیل و ارائۀ راهکارهای تقنینی برای مواجهه با تهدیدات مذکور است. نتیجه‌گیری مقاله، بازنگری و توسعۀ مقررات کیفری ایران با تأکید بر بزه‌انگاری‌های نو است تا ضمن تضمین عدالت کیفری و صیانت از امنیت ملی، امکان مدیریت تهدیدات نوظهور ناشی از هوش مصنوعی فراهم شود.
کلیدواژه‌ها
موضوعات

عنوان مقاله English

Criminalization of Emerging Threats Arising from Artificial Intelligence: An Analysis of the Challenges and Legislative Solutions

نویسندگان English

Saeed qomashi 1
Mohammadreza kameli 2
Mojtaba Nikokalam Nadim 3
1 Associate Professor, Department of Criminal Law and Criminology, Faculty of Law and Humanities, Kashan University, Kashan, Iran
2 M.A. in Criminal Law and Criminology, Faculty of Law and Humanities, Kashan University, Kashan, Iran (Corresponding Author)
3 M.A. in Criminal Law and Criminology, Faculty of Law and Humanities, Kashan University, Kashan, Iran
چکیده English

1. Introduction
The expansion of artificial intelligence has given rise to emerging threats and challenged fundamental concepts of criminal law, including mens rea (guilty mind), criminal intent, and the attribution of liability, because intelligent systems, in addition to facilitating the commission of offenses by humans, sometimes become sources of harmful conduct by virtue of their autonomous operation.
Deepfakes and intelligent phishing in the field of threats against persons threaten reputation and privacy, respectively, and facilitate access to information through behavioral data analysis. Errors in self-driving vehicles may lead to murder or bodily harm, and autonomous weapons increase the risk of violations of international humanitarian law. Artificial intelligence-based cyberattacks threaten critical infrastructure and national security; ransomware and the illicit mining of cryptocurrencies through botnets also damage property and trust in the digital economy.
Domestic studies, unlike some international research, have focused more on prevention; these include studies on the preventive capacity of artificial intelligence and Martha Stroppa's (2023) study on maintaining human control over autonomous cyber capabilities. The principal question is the adequacy of Iranian criminal law in addressing these threats. The hypothesis is that the laws are fragmented and inadequate, and the objective is to analyze the gaps and propose solutions in four areas: threats against persons, autonomous technologies, cybersecurity, and property.

2. Research Methodology
The research methodology is based on a theoretical-analytical approach and relies on documentary research and content analysis of domestic laws, international instruments, scholarly articles, specialized reports, and comparative studies. The findings are used to identify gaps and propose legislative solutions.

3. Research Findings
The findings of this study indicate that the existing criminal law provisions in Iran are fragmented and inadequate to address emerging threats arising from artificial intelligence. In the field of threats against persons, existing provisions, such as Articles 6 and 16 of the Computer Crimes Law, may be applicable to some instances of deepfakes and phishing, but these provisions do not provide for a distinct offense or a specific framework for such conduct.
In the field of autonomous systems and autonomous weapons, the Iranian criminal law system lacks a clear mechanism for determining criminal liability in cases where decision-making is independently performed by an algorithm. Artificial intelligence-based cyberattacks against critical infrastructure, despite their increasing speed and complexity, lack a distinct offense of “cyberattack” or a definition of “artificial intelligence-enhanced cybercrime.” In the field of property, automated ransomware and the illicit mining of cryptocurrencies through botnets call into question the adequacy of traditional classifications of financial crimes. Overall, the findings confirm the research hypothesis and demonstrate the need to reconsider the legal elements, criminal intent, and rules governing the attribution of liability.

4. Recommendations
To address legislative gaps, the article proposes draft statutory provisions for the criminalization of emerging technological conduct and the determination of criminal liability in autonomous systems:

1. Criminalization of Deepfakes in the Iranian Criminal Law System
Altering digital content to create a false and deceptive representation, with the intent to harm the personality, reputation, or pecuniary or non-pecuniary rights of natural or legal persons, shall constitute an offense. Where serious harm is caused, the maximum punishment shall apply; use for educational, research, artistic, or journalistic purposes, in the absence of any intent to deceive or cause harm, shall be excluded.

2. Criminalization of Phishing Attacks in Iranian Criminal Law
The use of fraudulent webpages, misleading emails, or text messages with the intent to commit fraud or unlawfully obtain information shall constitute an offense. The proposed punishments include taʿzir imprisonment of the sixth degree, confiscation of proceeds derived from the offense, and deprivation of the right to use computer and Internet services for a period of two to five years. Where serious harm occurs, the most severe punishment prescribed shall apply, and an unsuccessful attempt accompanied by criminal intent shall be deemed an attempt to commit an offense.

3. Criminalization of the Use of Artificial Intelligence-Based Autonomous Weapons
The use of autonomous weapons that violate international humanitarian law or the rules governing armed conflicts shall be prohibited and shall constitute an offense; participation in the production, design, distribution, sale, import, export, possession, or use of such weapons shall be subject to taʿzir imprisonment of the fifth degree, and in cases of murder or bodily harm, the corresponding punishments shall also apply. Where preventive measures are established and the offense occurs as a result of an unforeseeable defect, criminal liability shall be attributed to the system under the proposed provision. Government oversight, approval of high-risk modification, and the establishment of a compensation fund or insurance for compensation of losses is also proposed.

4. Criminalization of Conduct Arising from Artificial Intelligence in Self-Driving Vehicles
Where an error in the system is not attributable to any prior programming or planning by engineers or users and results in murder or bodily harm, the draft provides for the attribution of liability to the system and compensation for the damage caused; where it is not possible to directly enforce a punishment, liability shall, on a combined basis, rest with the owner, operator, or entity responsible for the system. Where there is a defect in design or manufacture, the manufacturing company or the owner shall be liable, and a special insurance fund is proposed.

5. Criminalization of Artificial Intelligence-Based Cyberattacks
Intelligent cyberattacks involving manipulation, destruction, unauthorized access to, or disruption of systems, critical infrastructure, or personal data shall be punishable by imprisonment and a pecuniary penalty of the fifth degree; serious harm to persons, national security, or the economy shall result in the maximum punishment. In an automated offense, the designer, programmer, owner, or operator shall be responsible, unless it is established that the commission of the offense resulted from factors beyond their control and foreseeability; lawful use is excluded.

6. Criminalization of Artificial Intelligence-Based Ransomware
The design, development, dissemination, distribution, or control of ransomware with the assistance of artificial intelligence, or the use of ransomware to threaten and extort natural or legal persons, shall constitute an offense. Serious damage to critical systems, national infrastructure, or public security shall result in the most severe punishment.

7. Criminalization of Cryptocurrency Mining Using Botnets and Artificial Intelligence
Cryptocurrency mining using artificial intelligence or botnets, without legal authorization from the competent authorities, shall be criminalized as unauthorized cryptocurrency mining.
At the legislative level, conduct that can be subsumed under existing offenses may be addressed through the amendment or addition of statutory provisions, including Articles 6 and 16 of the Computer Crimes Law concerning deepfakes; however, with regard to autonomous weapons and self-driving vehicles, an independent and comprehensive law is necessary because of the uncertainty surrounding the attribution of liability. The drafts should be completed through expert review and in accordance with international standards.

5. Conclusion
The existing criminal law frameworks in Iran lack the necessary coherence and adequacy to address emerging crimes based on artificial intelligence, particularly with regard to automated conduct and the attribution of liability. Traditional frameworks based on human reason, will, and intent have demonstrated their inadequacy in addressing autonomous technologies, and gaps are evident in defining offenses, identifying offenders, and establishing the mens rea (guilty mind).
The main innovation of this study lies in presenting a roadmap for four key areas, namely offenses against persons, security, property, and autonomous technologies: the independent criminalization of deepfakes and intelligent phishing, the definition of artificial intelligence-enhanced cybercrime, the establishment of specific regulations for ransomware and the illicit mining of cryptocurrencies, and the design of a combined liability regime for human actors and intelligent systems. The possibility of recognizing a special form of legal personality for artificial intelligence, based on financial autonomy, capacity to litigate, and layered liability, is also raised.
Effective counteraction requires clear legal definitions, criteria for establishing mens rea (guilty mind) in automated decisions, accuracy and safety standards, specialized evidentiary procedures, oversight, and international cooperation. An interdisciplinary approach should safeguard individual and collective rights while preserving the responsible development of artificial intelligence.

کلیدواژه‌ها English

Artificial intelligence
Criminalization
Deepfakes
Intelligent phishing
Legislative gaps in Iranian criminal law
دوره 2، شماره 3
بهار 1405
صفحه 1-53

  • تاریخ دریافت 23 مهر 1404
  • تاریخ بازنگری 01 اسفند 1404
  • تاریخ پذیرش 20 اسفند 1404
  • تاریخ اولین انتشار 01 فروردین 1405
  • تاریخ انتشار 01 فروردین 1405