Understanding the Artificial Intelligence (Ethics and Accountability) Bill, 2025
Introduction
In a significant move towards regulating artificial intelligence (AI) in India, Bharti Pardhi, a Member of Parliament from Madhya Pradesh, introduced the Artificial Intelligence (Ethics and Accountability) Bill, 2025, during the Winter Session of the Lok Sabha. This bill aims to create an ethics and accountability framework for the utilization of AI technologies, reflecting the growing recognition of the need for regulatory oversight amid AI’s rapid evolution.
Overview of the AI Ethics Bill
While the bill is a step forward, it falls short of addressing many critical AI-related issues, such as copyright concerns, rightful compensation, and clear licensing procedures. Instead, it primarily advocates for the establishment of an AI Ethics Committee composed of various experts tasked with creating ethical guidelines and ensuring compliance.
Proposal for an AI Ethics Committee
Defined in the bill, AI encompasses “computer systems or applications capable of performing tasks that typically require human intelligence.” The proposed Ethics Committee for AI will include:
- A Chairperson
- Representatives from various sectors—academia, industry, civil society, and government
- Experts appointed from fields such as law, data science, and human rights
This committee will be responsible for conducting regular audits to identify and mitigate algorithmic bias, ensuring that AI companies utilize diverse and inclusive training datasets while withdrawing biased systems until they are adequately corrected.
Responsibilities of the AI Ethics Committee
The bill outlines several key responsibilities for the committee, including:
- Developing and recommending ethical guidelines for AI technologies
- Monitoring compliance with established ethical standards
- Reviewing instances of misuse, bias, or violations of the act
- Promoting awareness and building capacity among relevant stakeholders
- Carrying out additional functions as prescribed
Developer Obligations Under the Bill
To ensure ethical compliance, the bill sets forth obligations for AI developers, requiring them to:
- Disclose their AI system’s purpose, limitations, and training data sources
- Conduct regular audits to prevent algorithmic bias and use inclusive training data
- Maintain records to demonstrate adherence to ethical standards
Moreover, the committee will have a grievance redressal mechanism and investigative powers, allowing users to lodge complaints regarding harm or misuse stemming from AI technologies.
Addressing Transparency & Accountability Issues
The bill emphasizes the rapid growth of AI technologies as a catalyst for economic and social progress but warns of potential ethical and regulatory pitfalls. Concerns raised include:
- Algorithmic bias
- Surveillance misuse
- Lack of transparency and accountability in automated decision-making
To counter these issues, the legislation aims to:
- Establish a comprehensive ethical deployment framework for AI
- Promote transparency, fairness, and accountability in AI design and implementation
- Safeguard individuals and communities from harm caused by unethical AI applications
- Encourage innovation and public trust through clear regulatory guidelines
Limiting AI-Based Surveillance
The bill proposes restrictions on the use of AI systems for surveillance, allowing such systems only for “lawful purposes” with prior approval from the Ethics Committee. However, the lack of definition for “lawful purposes” presents ambiguity, placing interpretative power in the hands of the committee.
Furthermore, AI applications in law enforcement, financial contexts, and employment must be free from discrimination based on race, religion, and gender, and are subject to rigorous ethical reviews.
AI in Indian Law Enforcement Context
Recent initiatives by various state governments showcase a trend towards leveraging AI for policing. For instance, the Maharashtra government has expanded MahaCrimeOS AI to all police stations, aiding in tasks like analyzing cybercrime complaints and drafting FIRs. Similarly, the Hyderabad Police is exploring AI’s potential in investigations, illustrating both the advantages and ethical complexities of AI surveillance.
Gaps in Copyright Regulation
Despite the bill’s contributions, one significant gap is its inadequate address of copyright issues. India currently lacks robust legislation regulating AI, and the government has shown reluctance to introduce new laws addressing AI-related concerns.
Key Considerations for Policymakers
Several crucial points merit attention from policymakers:
- The bill fails to establish a framework for ensuring fair compensation or royalties to data owners whose data is employed for AI model training.
- It does not clarify whether the committee can inspect datasets in response to complaints regarding outputs or data used.
- No specific processes are defined for addressing instances where compensation is lacking or deemed insufficient.
- The legislation overlooks accountability in copyright infringement, given that the existing framework permits AI training on publicly available data.
Conclusion
The introduction of the Artificial Intelligence (Ethics and Accountability) Bill, 2025, marks a pivotal moment in India’s journey towards responsible AI governance. While well-intentioned, it presents several challenges and omissions that necessitate further discussion and refinement. By addressing issues like copyright, compensation, and the ethical implications of AI applications, India can lay a strong foundation for a balanced approach to AI technologies that aligns with democratic principles and public interest.