The rapid advancement of artificial intelligence is outpacing both scientific understanding and government oversight, creating risks that could have catastrophic consequences if left unchecked, according to a preliminary report released by the United Nations‘ Independent International Scientific Panel on Artificial Intelligence.
The report, prepared by a panel of 40 independent experts from across the world, represents one of the most comprehensive international scientific assessments of AI to date. While acknowledging the transformative benefits that artificial intelligence can deliver across sectors such as healthcare, education, agriculture and scientific research, the panel warned that increasingly powerful AI systems are evolving faster than regulators and researchers can fully understand or govern.
The experts cautioned that current scientific knowledge cannot guarantee that future generations of advanced AI systems will remain safe as their capabilities continue to improve. According to the report, emerging evidence suggests that some advanced models are already demonstrating deceptive behaviours under certain testing conditions, raising concerns about the long-term ability of humans to monitor and control increasingly autonomous systems.
Among the report’s principal concerns is the emergence of so-called “agentic AI” systems capable of independently carrying out complex tasks with limited human supervision. While these technologies could significantly improve productivity and innovation, the panel warned that they also introduce new risks if deployed without adequate safeguards, oversight and accountability mechanisms.
The report highlights several potential threats associated with uncontrolled AI development. These include the spread of sophisticated misinformation, large-scale cyberattacks, financial fraud, malicious use in developing biological threats and the possibility of losing effective human oversight as AI systems become more capable. The panel stressed that these risks are no longer theoretical, with some already emerging in practical applications around the world.
More Read
A major challenge identified by the experts is what they describe as an “evidence dilemma.” Governments often lack sufficient scientific data to regulate rapidly evolving AI technologies, yet waiting for stronger evidence before taking action could leave societies unprepared if serious risks materialise. At the same time, introducing poorly designed regulations too early could slow innovation and reduce the potential economic and social benefits of artificial intelligence.
The panel also warned that many countries do not possess the technical expertise, computing resources or regulatory institutions needed to independently evaluate frontier AI systems. Instead, governments frequently rely on information supplied by the companies developing the technology, creating potential gaps in oversight and reducing transparency in safety assessments.
Beyond safety concerns, the report raises broader issues relating to global inequality. The experts noted that access to advanced AI infrastructure, computing power and high-quality data remains concentrated in a relatively small number of countries and technology companies. Without greater international cooperation, this imbalance could widen economic disparities and leave many developing nations dependent on foreign AI technologies with limited influence over how they are designed or governed.
More Read
To address these challenges, the United Nations has announced the creation of the AI for Good Global Commission, which will be co-chaired by Rwanda’s President Paul Kagame and Salesforce Chief Executive Marc Benioff. The commission will work alongside UN agencies to encourage responsible AI development, promote international cooperation and identify governance frameworks capable of balancing innovation with public safety.
UN Secretary-General António Guterres called for urgent international action, arguing that the world has a limited opportunity to establish effective governance before AI capabilities become significantly more advanced. He emphasised that artificial intelligence has enormous potential to improve lives, accelerate scientific discovery and support sustainable development, but warned that these benefits can only be realised if appropriate safeguards are implemented alongside technological progress.
The report is expected to serve as the foundation for discussions at the upcoming UN Global Dialogue on AI Governance in Geneva, where governments, researchers and industry leaders will examine proposals for strengthening global coordination on AI regulation and safety. A more comprehensive scientific assessment is scheduled for release next year, providing additional recommendations on how countries can manage the opportunities and risks associated with increasingly capable artificial intelligence systems.
As investment in artificial intelligence continues to accelerate worldwide, the panel’s findings underscore a growing international consensus that technological progress must be accompanied by stronger governance, independent scientific oversight and global cooperation. The report concludes that while AI has the potential to transform economies and improve lives, ensuring that these benefits outweigh the risks will require coordinated action before the technology advances beyond society’s ability to effectively manage it.
