I USE artificial intelligence (AI) regularly. I find it enormously useful, and I have little doubt that it will improve productivity, medicine, education, scientific research, and many other aspects of human life.
Technology has always changed society. But AI is different in an important respect. We are developing technology capable not merely of assisting human beings physically, but of generating knowledge, influencing decisions, profiling individuals, shaping opinions, and increasingly performing intellectual tasks previously undertaken by humans.
That raises questions which are not merely technological or commercial. They are questions of law, liberty, and human dignity.
The starting point should be a principle lawyers understand well: power requires limits.
We do not ordinarily accept that because something is commercially profitable or technologically possible, it should therefore be permitted without adequate regulation. Banks, pharmaceutical companies, employers, governments, and professionals operate within legal boundaries because their activities can affect the rights and welfare of others.
Why should immensely powerful AI systems be treated differently?
Our Federal Constitution protects fundamental liberties. Article 5 protects life and personal liberty, while Article 8 guarantees equality before the law and equal protection of the law. As AI becomes increasingly involved in employment, financial services, education, healthcare, policing, and government administration, difficult legal questions will inevitably arise about discrimination, accountability, privacy, procedural fairness and human autonomy.
Existing laws provide some protection. The Personal Data Protection Act 2010, for example, regulates the processing of personal data in commercial transactions. But AI presents questions extending considerably beyond conventional data protection.
Who is legally responsible when an automated system causes serious harm? Should a person be entitled to know when an important decision affecting him was substantially made by an algorithm? Should there be a right to meaningful human review? What safeguards should apply when AI is used by public authorities? At what point does algorithmic profiling become unlawful discrimination or an unacceptable intrusion into individual autonomy?
These are no longer science-fiction questions.
To its credit, Malaysia has recognised the challenge. On July 28, Prime Minister Datuk Seri Anwar Ibrahim launched AI Malaysia and the National AI Action Plan 2026-2030 as part of the broader aspiration of making Malaysia an “AI Nation” by 2030. I regard this as a positive and necessary development. The action plan itself recognises that technological progress must be balanced with trust, effective governance, and positive societal outcomes.
The government has also begun consultations on a proposed AI Governance Bill. This is another encouraging development. The important question now is what legal protections should accompany our national ambition.
Much of AI governance presently depends upon policies, standards, ethical principles, and voluntary guidance. These are important foundations. But principles such as fairness, privacy, transparency, accountability, and safety are truly tested when they conflict with commercial interests, administrative convenience, or the exercise of power.
That is where law becomes necessary.
I am not suggesting that every use of AI should be heavily regulated. That could suppress innovation and deprive Malaysia of enormous opportunities. Regulation should be proportionate to risk. An AI system recommending music plainly does not require the same scrutiny as one determining eligibility for employment, credit, medical treatment, insurance, or government benefits.
For high-risk applications, however, we should consider mandatory safeguards: impact assessments before deployment, transparency requirements, independent auditing where appropriate, protection against discriminatory outcomes, meaningful human oversight, and clear legal responsibility when harm occurs.
There is another dimension that troubles me even more.
For many years, through the Rapera effort, my friends and I have tried to encourage two qualities which I regard as essential to a healthy society: thinking and compassion.
AI makes those qualities more important, not less.
If we gradually outsource thinking to machines, allow algorithms to determine what information we see, permit technology to understand our preferences and vulnerabilities better than we understand them ourselves, and eventually become intellectually dependent upon systems controlled by governments or corporations, something profoundly human may be diminished.
The danger is therefore not simply that AI might become extraordinarily intelligent. It is that human beings might become less intellectually independent.
That is why this discussion cannot be left to computer scientists and technology companies. Lawyers, judges, parliamentarians, educators, philosophers, religious thinkers, parents, and ordinary citizens must participate.
If Malaysia genuinely intends to become an AI Nation by 2030, our ambition should not merely be to become proficient users and producers of AI. We should also aspire to become a society capable of governing it wisely.
The principle, to me, is simple: technology must remain the servant of humanity, not its master. Commercial progress must operate within law, and law must ultimately protect human dignity, freedom, and agency.
I do not believe that we can, or should, stop the development of AI. The real challenge is to ensure that as we embrace its enormous benefits, we put in place the laws and safeguards necessary to protect human freedom, dignity, and independence.
Malaysia has made a positive start. But AI is developing rapidly, and we must think seriously about where the boundaries should be before the harm is done.
Senior lawyer Datuk Seri Dr Jahaberdeen Mohamed Yunoos is the founder and chairman of Yayasan Rapera, an NGO that promotes community-based learning activities and compassionate thinking among Malaysians. The views expressed here are entirely his own.
