Experts from the United Nations have urged caution in the way artificial intelligence (AI) is discussed, emphasizing that the risks are more complex than often portrayed. During a recent event in New York, held alongside the United Nations General Assembly, several scientists and researchers warned against alarmist narratives that suggest AI could lead to human extinction. Joëlle Barral, a senior executive at Google DeepMind and a member of an independent international scientific committee on AI, noted that "investing in fear is not really effective," highlighting the need for a more measured approach to the topic. Some voices in the AI field have recently raised concerns about the dangers of the technology, with statements suggesting that AI could pose an existential threat to humanity within a decade. These claims were made by individuals like Jacob Coxon, a former employee of Anthropic and previously with OpenAI, and another research engineer who left Google DeepMind in July. Both suggested that the risks of AI might already be too great to avoid. However, experts argue that such dire predictions can oversimplify the issue and lead to unnecessary panic. Laurence Devillers, a professor and researcher in AI, emphasized that while the technology is significant, it is not inherently dangerous. She called for a focus on thoughtful reflection and human responsibility in managing AI's development. Maria Ressa, co-chair of the independent international scientific committee on AI, echoed this sentiment, saying that the conversation around AI risks has been too extreme. She pointed out that the reality is more nuanced, with future AI capabilities neither necessarily an existential threat nor completely inconsequential. The independent international scientific committee on AI was established in 2025, with its members selected in early 2026. Its goal is to deepen the scientific understanding of AI and inform United Nations discussions. Yoshua Bengio, one of the pioneers of AI and also a co-chair of the committee, stressed the importance of distinguishing between what is known and what is speculative. He noted that while it is crucial to discuss AI risks, doing so with factual accuracy and scientific rigor is more productive than allowing fear or ideology to dominate the conversation.