Senate bill seeks to regulate AI, proposes billion-rupee fines

• Bill seeks to mitigate "various risks" associated with AI.
• Violations could result in fines ranging from Rs1.5-Rs2.5bn.
• Digital rights activists raise concerns about proposed legislation.

Pakistan Muslim League-Nawaz (PML-N) Senator Afnan Ullah Khan has presented a bill in the Senate, seeking to regulate the use of Artificial Intelligence (AI) in the country to mitigate the "various risks" associated with the technology, imposing fines amounting to billions of rupees for violators.

The ruling-party senator, who presented the proposed 'Regulation of Artificial Intelligence Act 2024' as a private member bill, had tabled the draft on August 9 in the Senate, a The News report stated.

On September 9, the draft was referred to the Senate's standing committee on information technology for further deliberations. 

What are the objectives of the bill? 

In his statement of objectives, Senator Khan emphasised the "dire need to monitor, regulate, superintend, and supervise" the dissemination of data and information at both national and international levels. 

He highlighted the risks associated with AI use, calling for increased human involvement to prevent "an unseen catastrophe" and to establish a "secure and well-protected" AI system to drive growth and productivity in the country.

The bill further proposes the establishment of a "National Artificial Intelligence Commission" with its headquarters in Islamabad.

According to the bill, the commission would have several key mandates, including:

• Ensuring equitable access to AI opportunities for all citizens, regardless of religion, gender, ethnicity, geographic location, financial status, or physical ability.
• Maintaining human oversight over AI matters, particularly in areas such as national security, health, and education.
• Guaranteeing AI systems are safe, secure from "malicious use", and free from errors.
• Developing a robust AI research and development infrastructure.
• Utilising AI to enhance public service delivery, such as in healthcare, finance, trade, and science.

The bill also proposes protecting the privacy and safety of AI users and advocates for transparency in AI data collection, storage, and usage. Additionally, it calls for strengthening the country's telecommunications systems, digital governance, and "cybersecurity".

The commission members, who will be appointed by the government, will be prohibited from holding any public or political office simultaneously.

Moreover, violations of the bill could result in fines ranging from Rs1.5 billion to Rs2.5 billion, with the possibility of increased penalties in cases involving the "dissemination of sensitive or classified data". However, the bill does not specify which particular violations would trigger these fines.

In March, the United Nations General Assembly adopted its first global AI resolution, encouraging member states to monitor AI for risks to human rights, personal data, and fundamental freedoms.

What do digital experts think? 

Digital rights activists have raised concerns about the proposed legislation.

Nighat Dad, the founder of the Lahore-based Digital Rights Foundation, described the bill as "odd" and "abstract". 

"This bill talks about privacy [of personal data] but there is no data protection or privacy law in the country at the moment," she told The News, "This bill to me is just another piece of legislation that has been proposed without doing any ground work."

She also pointed out that the bill focuses heavily on forming a commission rather than addressing substantive AI-related issues.

Asad Baig, director of Media Matters for Democracy, highlighted that globally, one of the biggest debates surrounding AI involves the creation of deepfakes and potential solutions like watermarking AI-generated content.

"The bill doesn't specifically talk about this," he said. Baig added that, while AI has massive potential in the country's development, currently there is no need for a law to regulate or promote AI.

"What is needed instead is a cohesive implementation of AI-friendly policies in the sectors and academic research [on the topic] and additional funds to the higher education commission for research and development of AI-related projects," he added.

Mubariz Siddiqui, founder of Carbon Law, shared a similar view, arguing that the bill was not well thought-out. 

"Regulation of AI is complicated and something most governments in the world are working on," he said over the phone, "Any serious attempt to regulate AI would need to balance its potential misuses without stifling innovation in this sector."