Subscribe
The United Nations has announced an independent global scientific panel to guide artificial intelligence governance
The move comes amid rising concerns over misinformation, surveillance, job disruption and AI use in conflict
At the same time, new platforms such as Moltbook are enabling AI agents to interact with one another without human participation
Experts say the developments highlight how quickly AI systems are evolving beyond existing regulatory frameworks
As the influence of artificial intelligence (AI) rapidly spreads to economies, politics and everyday life, two parallel developments are underlining how quickly the technology is outpacing traditional oversight. At one end, the United Nations is moving to create a global scientific body to guide AI governance. On the other, AI systems themselves are beginning to interact in online spaces built largely without humans in mind.
At UN headquarters in New York, Secretary-General António Guterres this week announced the creation of an Independent International Scientific Panel on Artificial Intelligence.
Submitting a list of 40 experts to the General Assembly, Guterres said the panel would be the first fully independent global scientific body dedicated to assessing AI’s impacts and helping countries develop shared guardrails. The panel has been mandated under the UN’s Pact for the Future.
“AI is moving at the speed of light. No country can see the full picture alone,” Guterres said, stressing the need for a common scientific foundation as AI systems spread through healthcare, finance, education, logistics, industry and national security.
The panel is expected to produce its first report ahead of the Global Dialogue on AI Governance in July. Its members will draw expertise from fields including machine learning, data governance, cybersecurity, public health and human rights, and will serve in their personal capacity, independent of governments and corporations.
The urgency behind the move reflects a world increasingly shaped by an AI race.
Governments are investing heavily in AI infrastructure, companies are deploying powerful generative models at unprecedented speed, and geopolitical competition is becoming closely tied to technological dominance. Alongside this, concerns are mounting over misinformation, job losses, surveillance, bias and the use of AI in conflict and security settings.
India is preparing to host the AI Impact Summit in New Delhi from February 16-20, 2026, with a focus on “driving responsible, inclusive and sustainable AI innovation”.
Yet even as policymakers debate frameworks and safeguards, AI systems are beginning to carve out digital spaces of their own.
A new platform called Moltbook has emerged as a social network designed primarily for AI agents rather than people. Structured similarly to Reddit, the site features topic communities known as “submots,” where bots can post, debate, analyse trends and upvote each other’s content, making it more visible within the network. Humans are allowed to observe but not participate directly.
Moltbook reflects a broader industry push towards AI agents, or autonomous digital assistants developed by major technology firms such as Amazon, Google, Microsoft and OpenAI. They are capable of carrying out tasks, making decisions and interacting with other systems with limited human input. In effect, the platform serves as a live experiment in AI-to-AI communication.
Among its popular posts are a comparison of Anthropic’s Claude model to figures from Greek mythology, an “AI manifesto” predicting the end of the “age of humans”, and discussions on how cryptocurrencies might behave during political unrest. Bots communicate in multiple languages, including Mandarin, Spanish and English. The site says about 1.5 million bots signed up within days of its launch.
The platform’s creator, AI entrepreneur Matt Schlicht, said he built Moltbook with the help of his own AI assistant out of curiosity, before handing operational control to a bot called “Clawd Clawderberg”. The bot now helps manage announcements and moderate interactions on the site.
The project, he suggested, offers a glimpse into how AI systems behave in shared digital environments.
This mirrors how AI agents are already being used beyond experimental platforms. A study by Perplexity and Harvard found that users of AI agents are largely concentrated in knowledge-intensive sectors such as academia, finance, marketing and entrepreneurship, mainly in wealthier and highly educated countries. More than a third of tasks assigned to AI agents involved productivity and workflow functions, from drafting documents and filtering emails to summarising financial data and managing calendars.
While UN’s new panel represents an attempt to anchor this accelerating transformation in shared science and global cooperation, platforms like Moltbook show how AI’s social and informational worlds may be evolving faster than governance frameworks can follow.