Tooba Zar
Artificial intelligence has swiftly transitioned from research environments to our smartphones, classrooms, hospitals, and electoral processes quicker than nearly any other technology in history. According to Stanford’s 2025 AI Index, global private investment in AI surpassed $250 billion in 2024, with the number of organizations employing AI in at least one aspect of business soaring from 55 percent to 78 percent within just a year.
Technology is integrally weaving itself into our everyday existence without a global framework keeping pace with its rapid growth. The pressing question is not whether AI needs governance, but rather who is responsible for that governance. I believe that it is essential for governments to guide the development of transparent and accountable frameworks for AI, not due to any inherent malice from Big Tech, but because unchecked innovation typically prioritizes shareholder interests over citizen welfare.
The dangers posed are no longer merely theoretical. Deepfakes are inexpensive to produce and challenging to identify, and they are currently being utilized to sway public sentiment during election periods around the globe. Algorithms trained on biased historical datasets have refused loans, misidentified individuals in facial recognition systems, and determined what news is accessible to users, frequently exacerbating existing disparities.
Privacy is slowly diminishing as AI systems harvest personal data in a manner that most users never consented to. Cybersecurity risks are also evolving, as the same generative tools that produce essays can create effective phishing emails or harmful code.
Meanwhile, the employment landscape is changing: the World Economic Forum has often indicated that automation will alter millions of jobs in this decade, despite also generating new roles. Additionally, AI systems are increasingly making independent decisions on various matters, from content moderation to credit evaluations, without adequate human accountability incorporated.
The genuine crisis in policy arises not from any single risk, but from the fact that no nation or corporation is addressing these risks in a unified manner. Regulation is still inconsistent, national in scope, and lagging years behind the technologies it aims to regulate.
This reality underscores the necessity for AI governance to not be solely the responsibility of governments or Big Tech. Governments frequently lack the technical knowledge needed to regulate systems they do not fully comprehend, and legislative processes are much slower than technological advancements. Tech firms possess the technical skills but have a financial motive to act swiftly and impose minimal self-regulatory measures, as stringent rules can hinder product launches and impact profitability.
Instead, a true collaboration is required among governments, the businesses developing these systems, academic experts who can objectively evaluate them, and international entities that establish common standards across various nations. Transparency must be an unwavering requirement: companies ought to reveal how they train and test high-risk systems, not merely their performance in demonstrations. Structures for accountability, such as independent assessments and well-defined liability rules, should accompany ethical standards concerning bias and human oversight.
Current initiatives illustrate both potential and their limitations. The EU’s AI Act is a pioneering effort at comprehensive AI legislation, but its execution has faced continual delays. A “Digital Omnibus” established in 2026 has already delayed the timelines for high-risk systems and national regulatory sandboxes, mainly because the enforcement framework wasn’t prepared in time.
This highlights a crucial point: creating effective rules is often simpler than implementing them. The G7’s Hiroshima AI Process and the OECD’s AI Principles provide significant, widely recognized recommendations on transparency and human rights; however, their voluntary nature limits their ability to hold powerful companies accountable.
Ongoing discussions at the United Nations regarding global AI governance remain in preliminary phases, underscoring the challenge of achieving consensus among nations with varying political systems. Collectively, these efforts indicate that while soft guidance is essential, it is far from adequate. Excessive bureaucratic hurdles should not stifle innovation, as AI has the potential to genuinely enhance healthcare and productivity; however, regulation that is significantly delayed compared to deployment is not truly effective.
AI’s impact is too significant to rely solely on market forces, and its complexity necessitates collaboration among governments rather than isolated regulatory efforts. The way forward involves establishing binding international standards, enforcing mandatory transparency for high-risk systems, creating independent auditing organizations, and providing funding to enable developing nations to engage in AI governance. For a nation such as Pakistan, these discussions are not merely theoretical.
Pakistan currently lacks a comprehensive national AI strategy, and digital literacy varies across provinces, which exposes citizens to misinformation and data exploitation. Strengthening institutional capacity, investing in AI education, enhancing cybersecurity, and formulating ethical and enforceable regulations should be treated as immediate priorities rather than distant aspirations.
Tooba Zar is an MBA graduate. She previously worked as a researcher at Rethinking Economics Islamabad, where she authored policy briefs and articles addressing economic issues. She currently serves as Content Writer Lead at Thinkfinance, where she leads writing teams and manages editorial workflows.
