The IMF Finds a Third Word for AI

Key points

• The IMFC Chair’s Statement of 17 October 2026 refers to AI only as “advanced intelligence technologies”, after the United States pressed IMF members to adopt “super intelligence”, the label Executive Order 14434 requires US agencies to use.
• The IMF’s October 2026 World Economic Outlook builds no AI-driven productivity acceleration into its baseline; its AI scenarios run from a 0.6 per cent gain in world GDP by 2029 to a market correction that leaves global activity 1.1 per cent lower by 2028.
• In the same week the Financial Stability Board finalised 12 sound practices for AI adoption in finance, G7 finance ministers approved tabletop exercises on frontier-AI cyberattacks, and the IEA counted 134 data-centre grid policies in a survey of more than 34 jurisdictions.

When the International Monetary and Financial Committee (IMFC), the IMF’s ministerial steering body, closed its Bangkok meeting on 17 October 2026, its text discussed the technology that had dominated a week of forecasts without once calling it artificial intelligence. The Chair’s Statement of the Fifty-fourth Meeting of the IMFC (Press Release No. 26/340), issued by Saudi finance minister Mohammed Aljadaan, uses “advanced intelligence technologies” instead. Reuters reported that the final text made no reference to artificial intelligence, after the United States had pressed members to rename it “super intelligence”.

For the United States, the name of the technology is now a negotiating position. The dispute also landed in a week when international bodies moved from describing AI risk to building tools for it.

How a label from a US executive order reached an IMF text

The American word has a precise origin. Executive Order 14434, “Inaugurating the Era of Super Intelligence”, signed on 29 September 2026, instructs the executive branch to use “Super Intelligence” and “SI” in place of “Artificial Intelligence” and “AI”. It defines the new term by pointing back to the existing statutory definition of artificial intelligence at 15 U.S.C. § 9401(3), and gives the Assistant to the President for Science and Technology 60 days, to about 28 November, to propose legislative language for a federal definition.

On 15 October Erin Browne, the Treasury’s Under Secretary for International Affairs, took the label abroad. “President Trump has been really clear on this, and he’s, you know, using the term SI,” she told Reuters, in a report carried by the Oman Observer, adding: “Many countries are looking to the US to set the standards.” Reuters reported that Washington was negotiating with other shareholders over the IMFC text; a source told the agency that one possibility was to drop both terms in favour of “new technology”.

The result was a third term. Aljadaan told Reuters that all members agreed to the language. The statement warns that rising sovereign bond yields, together with uncertainties about the impact of advanced intelligence technologies on economic and financial issues, “could amplify macro-financial vulnerabilities”. Not every written submission to the same meeting adopted the new vocabulary: the IMFC statement submitted for the EU Council of Economic and Finance Ministers (Statement No. 54-4) speaks plainly of an “AI-driven investment cycle”.

AI & GPP has argued that global AI governance is splitting into two blocs around one technology. Bangkok suggests the split now reaches vocabulary, and that consensus bodies will absorb it by inventing neutral words.

The Fund prices AI as demand, not productivity

The World Economic Outlook, October 2026, Chapter 1 projects global growth of 3.1 per cent in 2026 and 3.4 per cent in 2027, and its baseline assumes no AI-driven acceleration in productivity growth, citing still-limited adoption. Box 1.2 models three AI scenarios against that baseline. In the first, a bigger build-out with early productivity gains “could increase world GDP by 0.6 percent by 2029”, with the biggest gains in the United States and Asia. In the second, Asian suppliers struggle to keep pace with the build-out and productivity gains arrive later, so world GDP is only 0.2 per cent higher by 2029. In the third, the productivity gains fail to materialise, AI investment unwinds and global activity ends up 1.1 per cent below the baseline by 2028, with deeper losses in the United States (1.8 per cent) and Asia (1.7 per cent). The modelled downside is almost twice the size of the upside. The chapter also warns of “self-reinforcing circular financing arrangements and increased leverage within the AI ecosystem”.

The Global Financial Stability Report launch on 13 October went further. Athanasios Vamvakidis of the IMF’s Monetary and Capital Markets Department named the first risk as “an AI-related market correction if expectations on investment or profits disappoint”. Banks, he said, remain resilient under a severe stress test combining AI, bond-market and energy shocks, but a tail of weak banks would end up under-capitalised in that scenario, and that tail has been growing.

Managing Director Kristalina Georgieva turned this into a case for rules. Speaking on an Annual Meetings panel on 16 October, she compared AI to a Formula 1 car, according to the Associated Press: “We are racing, and the question is how do we put barriers?” The aim, she said, was that “if a car was to fly out it doesn’t kill people”. Reuters, in a report published by Forbes México (in Spanish), said she called at a plenary session for urgent and practical international cooperation, and told the later panel she had lost sleep over how to regulate AI and keep poorer countries from falling behind. Browne, on a call with reporters, pushed back against those warnings: the Fund, she said, should take a more balanced approach and give equal weight to opportunities such as those driven by American innovation and superintelligence (translation by AI & GPP).

Supervisors moved from warnings to checklists and drills

On 15 October the Financial Stability Board finalised its sound practices for the responsible adoption of AI (Ref 14/2026): 12 practices in two groups, organisation-wide governance and lifecycle risk management. They run from board oversight and documented AI use to explainability, human oversight, AI-related cyber and ICT risk, and dependence on third-party AI providers. The FSB “strongly encourages” boards and senior management to reference them and calls them a “practical toolkit” firms can tailor. They are not binding, but supervisors now have a shared list to test against.

G7 finance ministers, meeting on the sidelines in Bangkok on 15 October, approved drills. “The G7 finance deputies had been discussing the idea of doing tabletop exercises on frontier AI,” Japan’s finance minister Satsuki Katayama said, Reuters reported. A Japanese official said G7 finance authorities would design simulated AI-driven cyberattacks, use the results to check current playbooks and rewrite manuals, and might involve private firms. No timetable or lead authority has been published.

The physical side arrived the day before. The International Energy Agency’s Powering Artificial Intelligence, published on 14 October at the request of France’s G7 presidency, surveys more than 34 policy jurisdictions covering about 90 per cent of global data-centre capacity and finds 134 policies and regulations on integrating data centres into electricity systems. Nearly 40 per cent of them date from the past 12 months. In Texas, Virginia, Finland and the United Kingdom, data-centre connection queues already exceed today’s peak demand, a pressure AI & GPP traced in Finland’s pause on Google’s groundwork. The report sets out six principles for decision-makers, including transparent tariff frameworks for large loads.

What decision-makers should do and watch

First, separate the label from the scope. Executive Order 14434 changes a word but, until a statutory definition is proposed, not the thing defined. Drafters of joint texts with the United States should anchor whatever term is agreed to an explicit definition, so that a label concession cannot later widen or narrow obligations.

Second, watch about 28 November. If the White House’s proposed federal definition of “super intelligence” departs from 15 U.S.C. § 9401(3), the vocabulary dispute becomes a scope dispute. Reuters notes that the United States holds the G20 presidency, so its finance-track texts are the next place to look. Consensus bodies elsewhere face similar tests: the autonomous-weapons talks covered in Consensus or Exit reach their review conference in Geneva in November.

Third, the IMF baseline contains no AI productivity dividend; budgets that assume one go beyond the Fund’s own forecast. Supervisors should map existing AI guidance against the FSB’s 12 practices now.

Questions this answers

What term did the IMF use for AI in October 2026? The IMFC Chair’s Statement of 17 October 2026 used “advanced intelligence technologies” and did not mention artificial intelligence, Reuters reported. The United States had pressed for “super intelligence”, the term required by Executive Order 14434.

Does the IMF’s October 2026 forecast assume AI will raise productivity? No. The World Economic Outlook’s Chapter 1 baseline assumes no AI-driven productivity acceleration. Its strongest AI scenario adds 0.6 per cent to world GDP by 2029, while a market-correction scenario leaves global activity 1.1 per cent below the baseline by 2028.

What did the FSB publish on AI in October 2026? On 15 October the Financial Stability Board finalised 12 sound practices for responsible AI adoption by financial institutions, covering governance and lifecycle risk management. They are not binding.

Sources

1. Chair’s Statement of the Fifty-fourth Meeting of the International Monetary and Financial Committee (IMFC), Press Release No. 26/340 — International Monetary Fund, 17 October 2026
2. IMF members call for more support amid war shocks, ‘advanced intelligence’ risks — ThePrint (Reuters), 17 October 2026
3. Inaugurating the Era of Super Intelligence (Executive Order 14434) — The White House, 29 September 2026
4. US Treasury urges renaming AI as ‘super intelligence’ — Oman Observer (Reuters), 15 October 2026
5. IMFC Statement No. 54-4: Statement by Mr. Harris, EU Council of Economic and Finance Ministers — International Monetary Fund, Fifty-fourth IMFC meeting, 16–17 October 2026 (PDF)
6. World Economic Outlook, October 2026, Chapter 1: Global Prospects and Policies — International Monetary Fund, October 2026 (PDF)
7. IMF / Global Financial Stability Report Annual Meetings, 2026 — IMF Media Center, 13 October 2026 (video briefing)
8. IMF-World Bank meetings focus on global risks and potential benefits of AI — Associated Press via WSLS, 16 October 2026
9. FMI insta a cooperar en IA y riesgos de guerra, mientras EU celebra la economía — Forbes México (Reuters), 16 October 2026 (in Spanish; translation by AI & GPP)
10. FSB finalises sound practices for the responsible adoption of artificial intelligence (AI) — Financial Stability Board, 15 October 2026
11. G7 to launch exercises to guard against AI cyberattacks — TimesLIVE (Reuters), 16 October 2026
12. Powering Artificial Intelligence: Executive summary — International Energy Agency, 14 October 2026