Just a few years ago, AI systems were limited to answering questions or generating basic text. Today, they can write computer code, analyse large datasets, produce realistic images and videos, assist in scientific discovery and drug development, and increasingly operate with minimal human oversight.
However, while AI capabilities are advancing rapidly, experts warn that the regulatory frameworks meant to ensure its safe use are not keeping pace.
This is the key finding of a preliminary report by the UN Independent International Scientific Panel on Artificial Intelligence, released on Wednesday. The report cautions that the opportunity to establish effective global governance is still open, but may not remain so for long, as AI continues to emerge as one of humanity’s most transformative technologies.
If used responsibly, AI could significantly accelerate progress toward the Sustainable Development Goals by enhancing healthcare, education, scientific research, agriculture, and accessibility for people with disabilities.
At the same time, without proper safeguards, it could worsen inequality, amplify misinformation, undermine human rights, disrupt labour markets, and concentrate powerful capabilities in the hands of a few governments and corporations.
The central challenge, the report notes, is balancing the immense benefits of AI with its growing risks. Over recent years, AI systems have advanced at an extraordinary pace, driven by powerful computing infrastructure, vast datasets, and improved algorithms. These developments have enabled fluent conversational systems, advanced reasoning tools, software engineering capabilities, and highly realistic multimedia generation. A new wave of even more capable systems is already emerging.
The report also highlights growing real-world benefits, including breakthroughs in healthcare, improved food security, and broader social impact.
However, the same technologies are also giving rise to new risks, including online harm, disinformation, criminal misuse, mental health concerns, loss of control, and environmental costs.
The benefits of AI remain unevenly distributed. Although adoption is global, access to the infrastructure, expertise, data, and investment needed to fully harness AI remains concentrated in developed countries.
Many developing nations lack the resources to build, evaluate, or adapt these systems for local needs, leaving them dependent on technologies they cannot fully control or audit.
The panel warns that if these disparities persist, AI could deepen existing global inequalities rather than reduce them.
While more than 40 AI governance frameworks and ethical guidelines already exist worldwide, they remain fragmented, inconsistent, and largely untested for effectiveness.
The panel’s findings will inform the UN Global Dialogue on AI Governance, set to begin in Geneva on 6 July 2026, where member states will discuss coordinated international approaches to AI regulation.
The report concludes that AI itself is neither inherently beneficial nor harmful. Its ultimate impact will depend on the decisions made today by governments, companies, and societies.
AI is already reshaping science, healthcare, education, and economies worldwide.
