A Call for Calm from Global Experts
The message comes from the UN's High-Level Advisory Body on AI, a diverse group of 39 experts from across the globe tasked with recommending a path for international governance. In a series of recent reports and dialogues, this body has warned that AI is advancing
at a 'runaway speed,' often outpacing the ability of institutions to manage it. Their central plea is for governments to resist the urge to create laws based on speculative hype or worst-case scenarios. Instead, they champion an approach grounded in solid, scientific evidence to navigate both the immense opportunities and the tangible risks of AI, from its potential to accelerate the Sustainable Development Goals to threats of misinformation and bias.
What 'Evidence-Based' AI Policy Means
Evidence-based policymaking means replacing intuition and political pressure with decisions grounded in data and objective analysis. In the context of AI, this involves several concrete steps. Experts suggest incentivising independent evaluations of AI models before they are released to the public and requiring companies to be more transparent about their safety practices. It also means establishing 'regulatory sandboxes' — controlled environments where new AI technologies can be tested to understand their real-world impact. This approach focuses on addressing measurable harms, such as algorithmic bias or privacy violations, rather than reacting to far-off, hypothetical dangers. The goal is to create rules that are flexible enough to adapt as the technology itself evolves, preventing today's laws from becoming quickly outdated.
The Dangers of Flying Blind
Crafting policy based on guesswork carries significant risks. Premature or poorly designed regulations can stifle innovation, creating massive compliance costs that disproportionately harm smaller startups and concentrate power in the hands of large tech firms. A fragmented patchwork of rules, where different states or countries have conflicting laws, can create legal uncertainty and make it impossible for beneficial technologies to operate across borders. This could slow the development of AI systems with life-saving potential, such as those used in medical diagnostics or for improving vehicle safety. Experts warn that focusing on speculative threats can also divert attention and resources away from addressing the real and present harms AI can cause, such as enabling sophisticated fraud, amplifying discrimination, or its significant environmental footprint.
A Global Quest for Common Ground
The UN panel's call for an evidence-based approach comes as nations pursue vastly different regulatory strategies. The European Union has adopted the comprehensive 'AI Act,' a broad legal framework, while the United States has so far favoured a more market-driven approach. India, with its massive digital ecosystem and ambitions to be a global AI leader, is navigating its own path, balancing innovation with the need for guardrails. The UN's initiative is an attempt to find a globally inclusive and interoperable framework. The advisory body has identified a global governance deficit, noting that a patchwork of norms is emerging where many countries are excluded from key conversations despite being affected by AI. The recent Global Dialogue on AI Governance in Geneva is a key step in bringing every country to the table to build a shared understanding based on scientific consensus.
