Artificial intelligence has been integrated into applications across diverse sectors, from automobiles to agriculture. According to the Grand View Research report, the AI market is projected to grow at a CAGR of 36.6% from 2024 to 2030. However, with the incorporation of these rapid innovations, it’s equally essential to address the safe and ethical use of AI.
This is where the need for artificial intelligence regulations comes into the picture. Without regulation, AI can lead to issues such as social discrimination, national security risks, and other significant issues. Let’s look into the details of why artificial intelligence regulations are necessary and what you need to understand about them.
Why AI Regulation is Required?
The progressively increasing use of AI in various domains globally has brought with it certain challenges. This has led to the need for regulatory frameworks. Here are some of the critical reasons why AI regulation is essential:
Threat to Data Privacy
The models and algorithms in AI applications are trained on massive datasets. These datasets contain data records consisting of personally identifiable information, biometric data, location, or financial data.
To protect such sensitive data, you can deploy data governance and security measures at the organizational level. However, these mechanisms alone cannot ensure data protection.
Setting guidelines at the regional or global level to obtain consent from people before using their data for AI purposes can ensure better data protection. This facilitates the preservation of individual rights and also establishes a common consensus among all stakeholders on using AI.
Ethical Concerns
If the training datasets of AI models contain biased or discriminatory data, it will reflect in the outcomes of AI applications. Without proper regulations, such biases can affect decisions in hiring, lending, or insurance issuance processes. The absence of guidelines for using artificial intelligence in judiciary proceedings can lead to discriminatory judgments and erosion of public trust in the law.
Regulatory frameworks compelling regular audits of AI models could be an efficient way to address ethical issues in AI. Having a benchmark for data quality and collecting data from diverse sources enables you to prepare an inclusive dataset.
Lack of Accountability
If there is an instance of misuse of an AI system like deepfakes, it can be difficult to impart justice to the victims. This is because without a regulatory framework, no specific stakeholder can be held responsible. Having a robust set of artificial intelligence regulations helps resolve this issue by clearly defining the roles of all stakeholders involved in AI deployment.
With such an arrangement, developers, users, deployers, and any other entity involved can be held accountable for any mishaps. To foster transparency, regulatory frameworks should also make it compulsory to document the training process of AI models and how they make any specific decision.
Important AI Regulatory Frameworks Around the World
Let’s discuss some artificial intelligence laws enforced by various countries around the world:
India
Currently, India lacks specific codified laws that regulate artificial intelligence. However, some frameworks and guidelines were developed in the past few years to introduce a few directives:
- Digital Data Protection Act, 2023, which is yet to be enforced to manage personal data.
- Principles for Responsible AI, February 2021, contains provisions for ethical deployment of AI across different sectors.
- Operationalizing Principles for Responsible AI, August 2021, emphasized the need for regulatory policies and capacity building for using AI.
- National Strategy for Artificial Intelligence, June 2018, was framed to build robust AI regulations in the future.
A draft of the National Data Governance Framework Policy was also introduced in May 2022. It is intended to streamline data collection and management practices to provide a suitable ecosystem for AI-driven research and startups.
To further promote the use of AI, the Ministry of Electronics and Information Technology (MeitY) has created a committee that regularly develops reports on development and safety concerns related to AI.
EU
The European Union (EU), an organization of 27 European countries, has framed the EU AI Act to govern the use of AI in Europe. Adopted in March 2024, it is the world’s first comprehensive law on artificial intelligence regulations.
While framing the law, different applications of AI were analyzed and categorized according to the risks involved in them. The Act categorizes AI applications based on risk levels:
- Unacceptable Risk: This includes applications like cognitive behavioral manipulation and social scoring. Real-time remote biometric identification is permitted under stringent conditions.
- High Risk: This includes AI systems that negatively impact people’s safety or fundamental rights. Under this category, services using AI in the management of critical infrastructure, education, employment, and law enforcement have to register in the EU database.
To further ensure safety, the law directs Generative AI applications such as ChatGPT to follow the transparency norms and EU copyright law. More advanced AI models, such as GPT-4, are monitored, and any serious incident is reported to the European Commission.
To address issues such as deepfake, this AI law has made provisions that AI-generated content involving images, audio, or video should be clearly labeled.
The intent of the EU AI Act is to promote safe, transparent, non-discriminatory, ethical, and environment-friendly use of AI. The Act also directs national authorities to provide a conducive environment for companies to test AI models before public deployment.
USA
Currently, there is no comprehensive law in the USA that monitors AI development, but several federal laws address AI-related concerns. In 2022, the US administration proposed a blueprint for an AI Bill of Rights. It was drafted by the White House Office of Science and Technology (OSTP) in collaboration with human rights groups and common people. The OSTP also took input from companies such as Microsoft and Google.
The AI Bill of Rights aims to address AI challenges by building safe systems and avoiding algorithmic discrimination. It has provisions for protecting data privacy and issuing notices explaining AI decisions for transparent usage. The bill also necessitates human interventions in AI operations.
Earlier, the US issued some AI guidelines, such as the Executive Order on the Safe, Secure, and Trustworthy Use of Artificial Intelligence. It requires AI developers to report potentially threatening outcomes to national security.
Apart from this, the Department of Defense, the US Agency for International Development, and the Equal Employment Opportunity Commission have also issued orders for the ethical use of AI. Industry-specific bodies like the Federal Trade Commission, the US Copyright Office, and the Food and Drug Administration have implemented regulations for ethical AI use.
China
China has different AI regulatory laws at national, regional, and local levels. Its Deep Synthesis Provisions monitors deepfake content, emphasizing content labeling, data security, and personal information protection.
The Internet Information Service Algorithmic Recommendation Management Provisions mandates AI-based personalized recommendation providers to protect your rights. These provisions are grouped as general provisions, informed service norms, and user rights protection. It includes directions to protect the identity of minors and allows you to delete tags about your personal characteristics.
To regulate Generative AI applications, China recently came up with interim measures on generative AI. It directs GenAI service providers that they should not endanger China’s national security or promote ethnic discrimination.
To strengthen the responsible use of AI, China has also deployed the Personal Information Protection Law, the New Generation AI Ethics Specification, and the Shanghai Regulations.
Several other nations, including Canada, South Korea, Australia, and Japan, are also taking proactive measures to regulate AI for ethical use.
Challenges in Regulating AI
The regulation of AI brings with it several challenges owing to its rapid evolution and complex nature. Here are some of the notable challenges:
Defining AI
There are varied views regarding the definition of artificial intelligence. It is a broad term that involves the use of diverse technologies, including machine learning, robotics, and computer vision. As a result, it becomes difficult to establish a one-size-fits-all regulatory framework. For example, you cannot monitor AI systems like chatbots, automated vehicles, or AI-powered medical diagnostic tools with the same set of regulations.
Cross-Border Consensus
Different regions and nations, such as the EU, China, the US, and India, are adopting different regulations for AI. For example, the AI guidelines of the EU emphasize transparency, while those of the US focus on innovation. Such an approach creates operational bottlenecks in a globalized market, complicating compliance for multinational entities.
Balancing Innovation and Regulation
Excessive AI regulation can hamper the development of AI to the full extent, while under-regulation can lead to ethical breaches and security issues. Most companies avoid implementing too many regulations, fearing that it could reduce innovation.
Rapid Pace of Development
The speed with which AI is developing is outpacing the rate at which regulations are developed and enforced. For instance, considerable damage occurred even before regulatory bodies could create rules against deepfake technology. It is also challenging for regulators to create long-term guidelines that can adapt to the rapidly evolving nature of AI technologies.
Lack of Expertise among Policymakers
Effective AI regulation requires policymakers to have a good understanding of the potential risks and mitigation strategies of this technology. Policymakers usually lack this expertise, leading to the designing of policies that are irrelevant or insufficient for proper monitoring of AI usage.
Key Components of Effective AI Regulation
Here are a few components that are essential to overcome hurdles in framing artificial intelligence regulations:
Data Protection
AI systems are highly data-dependent, which makes it crucial for you to prevent data misuse or mishandling. Regulations like GDPR or HIPAA ensure that personal data is utilized with consent and responsibly.
You can take measures such as limiting data retention time, masking data wherever required, and empowering individuals to control how their personal information is used.
Transparency
AI systems often operate as black boxes that are difficult to understand. Having transparency ensures that the processes behind AI decision-making are accessible for verification.
To achieve this, the regulatory framework mandates companies to design AI products with auditing features so that the underlying decision-making logic is accessible for verification. If there are any discrepancies, you can challenge the AI decisions, and the developers will be held accountable for providing remedies.
Human Oversight
A fully autonomous AI system makes all decisions on its own and can sometimes take actions that lead to undesirable consequences. As a result, it is important to have some proportion of human intervention, especially in sectors such as healthcare, finance, and national security.
For this, you can set up override mechanisms where humans can immediately intervene when AI behaves irrationally or unexpectedly.
Global Standards and Interoperability
With the increase in cross-border transactions, it is essential to develop AI systems that facilitate interoperability and adhere to global standards. This will simplify cross-border operations, promote international collaboration, and reduce legal disputes over AI technologies.
Way Forward
There has been an increase in instances of misuse of AI, including deepfakes, impersonations, and data breaches. Given this trend, artificial intelligence regulations have become the need of the hour.
Several countries have introduced legal frameworks at basic or advanced levels to address many artificial intelligence-related questions. However, we still need to fully understand the implications AI will have on human lives.
In the meantime, it is the responsibility of policymakers and technical experts to create public awareness about the impacts of good and bad use of AI. This can be done through education, training, and public engagement through digital platforms. With these initiatives, we can ensure that AI’s positive aspects overpower its negative consequences.
FAQs
To whom does the EU AI Act apply?
The EU AI Act applies to all businesses operating within the EU. It is compulsory for providers, deployers, distributors, importers, and producers of AI systems to abide by the rules of the EU AI Act.
What are some other examples of data protection regulations?
Some popular examples of the data protection regulations include:
- Digital Personal Data Protection (DPDP) Act, India
- General Data Protection Regulation (GDPR), EU
- California Consumer Privacy Act (CCPA), US
- Protection of Personal Information Act (POPIA), South Africa
- Personal Information Protection and Electronic Documents Act (PIPEDA), Canada