ICC Russia Secretary General Tatiana Monaghan is in New York, taking part in International Chamber of Commerce events during the high-level week of the 81st session of the UN General Assembly.
Artificial intelligence has become one of the week's central themes: the discussion is shifting from technological development towards risks to security and competition.
Two approaches to global AI governance
UN Secretary-General António Guterres called for the responsible development of AI and stressed that decisions over life and death must not be delegated to machines. US President Donald Trump stated that Washington rejects attempts to create a global system of control over AI.
Security Council: risks through the eyes of developers
On 23 September, at France's initiative, the Security Council held a briefing on ‘Artificial Intelligence and International Security’. Speakers included Yoshua Bengio, co-chair of the UN's Independent International Scientific Panel on AI, and the heads of OpenAI, Anthropic and Hugging Face.
Bengio: described the risks of frontier systems as “real and imminent” and proposed independent assessment and licensing modelled on medicine, aviation and nuclear energy.
Dario Amodei (Anthropic): under poor governance, AI could become a risk to humanity as a whole.
Sam Altman (OpenAI): the most important decisions cannot be made by laboratories in San Francisco alone; comparable international standards for risk assessment are needed.
Clément Delangue (Hugging Face): spoke against slowing development but in favour of disclosing incidents and open models, warning of the danger of concentrating computing resources in the hands of a few companies and states.
State positions
United States: the risks are not a reason to “halt further development” of AI or restrict it with “new global governance structures”.
China: warns against dividing the world into technological camps on an “us versus them” basis and spoke in favour of open models.
France: calls for common rules to be developed for frontier AI systems.
Pakistan: security measures must not become a tool of technological exclusion; decisions on AI risks cannot remain in the hands of “a handful of companies and capitals”.
United Kingdom: intends to make AI one of the themes of its G20 presidency in 2027.
The Council adopted no decisions.
Energy: the material side of the AI race
France drew attention to the growing energy and water consumption of data centres and to the competition for these resources with other industries. According to estimates by the International Energy Agency, cited by UNECE, global electricity consumption by such centres could nearly double — from 485 TWh in 2025 to 950 TWh by 2030, around 3% of global demand.
In our view, competitiveness in AI will increasingly depend not only on models and data, but also on the availability of electricity and grid infrastructure. For countries with well-developed energy sectors, this is a distinct strategic advantage.
What is next?
On 24 September, Donald Trump and Xi Jinping are due to meet in Washington; according to Reuters, the two sides are discussing a dialogue on AI safety, including the exchange of information on incidents. On 26 September, Sergey Lavrov is due to address the General Assembly. The next session of the UN Global Dialogue on AI Governance will take place in New York in 2027.
For business
At the international level, proposals were put forward for the independent assessment of powerful AI systems, notification of serious incidents, developer liability and risk insurance. These are not yet binding rules, but an emerging agenda.
IMPORTANT! Companies using AI in foreign trade operations, settlements and trade finance would be well advised to put internal procedures in place now: identify where AI is used, retain human control over decisions, assess the risk of errors, and record them.