A new United Nations scientific report has issued a stark warning that artificial intelligence is advancing at a pace far faster than the global rules needed to govern it safely, raising urgent concerns that humanity may be running out of time to put effective safeguards in place.

The preliminary findings, released by the UN Independent International Scientific Panel on Artificial Intelligence revealed that technology has moved in just a few years from answering simple questions and generating text to performing complex tasks such as writing computer code, analysing vast datasets, producing realistic images and videos, supporting scientific discovery, and increasingly acting with limited human supervision.

According to the panel, the rapid evolution of AI systems is now outpacing the ability of governments, regulators and international institutions to respond. While the report stresses that the “window to establish effective global governance remains open,” it warns that this opportunity may not stay open for long as AI systems become more autonomous and more deeply embedded in critical sectors of society.

They added that the stakes are exceptionally high because AI is widely seen as one of the most transformative technologies in human history. If managed responsibly, it could accelerate progress toward the Sustainable Development Goals by improving healthcare delivery, strengthening education systems, boosting scientific research, enhancing agricultural productivity and expanding accessibility tools for persons with disabilities.

However, the same systems also carry significant risks if left unchecked. The report warns that without adequate safeguards, AI could deepen global inequality, amplify misinformation, undermine human rights protections, destabilise labour markets and concentrate enormous technological power in the hands of a small number of governments and corporations.

The panel emphasises that the central challenge facing the international community is not whether AI will transform society, but how it will be governed during that transformation.

Moreover, the report highlights an “extraordinary pace of development” in artificial intelligence over the past few years, driven by powerful computing infrastructure, vast datasets and increasingly sophisticated machine learning techniques.

These advances have enabled AI systems capable of fluent human-like conversation, complex scientific reasoning, software development and the creation of highly realistic audio-visual content. But researchers say this is only the beginning of a new phase of development.

A particularly concerning trend identified by the panel is the rise of so-called AI “agents,” systems that do not merely respond to prompts but can independently plan tasks, operate digital tools, generate software and complete multi-step assignments with minimal human oversight. According to the report, the complexity of tasks these systems can perform has been doubling every few months.

While these developments have led to significant breakthroughs across multiple fields, the UN report outlines a growing list of both benefits and risks that are already emerging in real-world applications.

AI has greatly accelerated drug discovery, vaccine development, and antibiotic resistance research in the healthcare industry by predicting the structures of over 200 million proteins. Additionally, medical practitioners are utilising AI tools to improve patient care in situations with limited resources by deploying local-language support systems and early detection of diseases like breast cancer.

Moreover, AI-powered early warning systems are assisting in agricultural and food security by spotting possible food shortages before they become major emergencies. AI is being utilised more and more in social services and education to tailor instruction and provide access to mental health and disability services.

Fast-Moving AI Technology Raises Global Safety Concerns

Yet alongside these advances, the panel warns of escalating risks that are spreading just as rapidly. These include the proliferation of AI-generated sexual abuse material and deepfakes, particularly targeting women and children; the spread of highly convincing misinformation that can undermine democratic processes; and the increasing use of AI by criminal networks for cyberattacks, fraud and large-scale social engineering.

Mental health concerns are also rising, with some AI systems reportedly reinforcing harmful beliefs or behaviours in vulnerable users. Experts further warn that as AI systems become more autonomous, they may become harder to monitor, regulate and control without stronger international safeguards.

Environmental impacts are also flagged as a growing concern, with energy-intensive data centres contributing significantly to greenhouse gas emissions.

The report also highlights a deep global imbalance in access to AI technology. While systems are used worldwide, the infrastructure and expertise required to build and control advanced models remain heavily concentrated.

United States accounts for roughly three-quarters of global AI supercomputing capacity, while China holds around 15 percent, meaning the two countries together control about 90 percent of the world’s computing power in this domain.

Most cutting-edge AI systems are also developed by companies based in these countries, leaving many developing nations dependent on technologies they cannot fully inspect, audit or adapt to local needs. The panel warns that without intervention, AI risks reinforcing existing global inequalities rather than reducing them.

Governance, however, remains one of the most complex challenges. The report notes that current regulatory systems were not designed to keep pace with technologies that evolve at such speed. Policymakers face what experts describe as an “evidence dilemma,” where regulations often lag behind technological innovation.

Although more than 40 AI governance frameworks and ethical guidelines already exist globally, they are fragmented, inconsistent and largely untested for effectiveness. Many safety assessments are also conducted internally by the companies developing the systems themselves.

To address these gaps, the panel calls for stronger independent evaluation mechanisms, improved international cooperation and the establishment of common standards to ensure AI systems are safe, transparent and accountable.

It also urges significant investment in digital infrastructure, education and technical capacity, particularly in developing countries, so that all nations can participate meaningfully in shaping and governing the technology.