ARTIFICIAL intelligence could revolutionise Pakistan’s legal system, as recognised by Judge Mohammad Munir of the Mandi Bahauddin district court in his detailed order, in a case of pre-arrest bail for a juvenile. The case was decided with the aid of ChatGPT 4. The technology can benefit legal practitioners and judges alike in identifying patterns and trends that may otherwise go unnoticed.

AI offers an attractive solution to our backlog of cases through its potential to improve the efficiency of decision-making by substantially reducing the judges’ workload. However, amidst the promise of efficiency and speed are concerns that require the judiciary to tread the road of automation with caution. Various ethical, legal and societal implications of employing AI require consideration, which are particularly relevant in the context of Pakistan where the legal system is still evolving.

AI systems are as unbiased as the data they are trained on. If the data contains biases, the system will perpetuate them. This is particularly concerning as the current state of Pakistan’s laws and judicial precedents is strewn with prejudices against certain sections of the population such as minorities and women. Imagine AI deciding cases solely on the basis of available data in a country where the two-finger test in rape cases was abolished just two years ago. True, AI-powered decision-making can bring consistency to Pakistan’s judicial system. However, ‘stare decisis’ (following precedent) is not the only criterion for justice; rather, the law should be developed/altered empathetically and according to society’s changing needs. Slavish adherence to existing data that AI offers in the current unregulated framework runs the risk of dehumanising the system and perpetuating biases.

If the data contains biases, AI will perpetuate them.

Secondly, AI-powered systems can be difficult to understand and explain, even by experts. This is not to take anything away from Mr Munir, who only used ChatGPT 4 as an inspiration to write the judgement without substituting it with his role as judge. However, not all judges may be expected to use it with the same care unless provided with guidelines. In the absence of regulation and clear guidelines, the lack of transparency in AI’s use makes it difficult to determine whether the decisions being made by the system are fair. Transparency is key to maintaining public trust and thus, it is imperative that AI be introduced in a way that ensures transparency given the deep public distrust of the judicial system.

Moreover, Pakistan’s legal system is complex and riddled with uncertainties, which raises concerns about accountability. If an AI system makes a mistake, who is responsible for the decision? AI’s introduction without carefully considered regulation would exacerbate the issue of holding individuals accountable. Additionally, the fundamental principles of due process and the right to fair trial may be compromised in a system where decisions are made without human input.

Another obvious ethical concern revolves around data privacy and increased surveillance. The current legal framework of data privacy and surveillance contained mainly in the Pakistan Electronic Crimes Act, 2019, and the Pakistan Telecommunication Authority (Re-Organisation) Act, 1996, is already under criticism for threatening civil liberties, its vagueness and for not adequately protecting citizens’ data.

Any AI-system integrated into Pakistan’s judicial system would have to be trained on and given access to large amounts of personal data in order to make legal decisions — which could lead to further violation of privacy. Moreover, we are currently unaware of the complete extent of the utility and potential of AI in courts — regardless of its potential to be used to monitor and track individuals in courtrooms, their movements, behaviour and speech patterns to predict criminal behaviour or evaluate the risk of recidivism. This could lead to increased surveillance and control of individuals. It is crucial to ensure that AI is integrated in a way that respects individual rights and freedoms.

The concerns raised above are not to suggest that AI must not be integrated into the legal system. Rather, if we are to wield AI in courts without a hitch, we must first regulate, illuminate, and empathise its use. As Alan Kay put it: “AI in courtrooms must be balanced with the fundamental principles of justice, ensuring fairness, transparency and accountability.”

The leap towards an AI-powered judicial system requires that the AI system introduced is developed and trained on unbiased data, and well-considered regulations are put in place which ensure transparency and accountability as well as recognise the limitations of using AI in courts.

The writer is a lawyer and president of the technology Law Research Society.
Twitter: @muhammadareebk

Published in Dawn, April 26th, 2023