The world’s largest sovereign wealth fund has begun using artificial intelligence to screen investments for ethical and reputational risks, marking a significant shift in how global investors integrate technology into decision-making and sustainability oversight.
Norges Bank Investment Management (NBIM), which manages Norway’s US$2.2 trillion oil fund, said Thursday it is deploying advanced artificial intelligence tools to analyse companies entering its vast investment portfolio, strengthening its environmental, social and governance (ESG) monitoring framework.
The fund built from Norway’s oil and gas revenues holds stakes in more than 7,200 companies across 60 countries and owns roughly 1.5 percent of all publicly listed equities worldwide, making its investment decisions closely watched by markets and policymakers.
NBIM confirmed that its ESG risk monitoring team has been using the Claude AI model developed by Anthropic since late 2024 to identify governance, sustainability and reputational risks linked to potential investments.
According to the fund’s annual responsible investment report, artificial intelligence now scans companies immediately after they enter the equity portfolio, enabling analysts to flag concerns such as corruption exposure, forced labour risks or fraud allegations within 24 hours.
“These tools help us rapidly scan a wide range of public information that goes beyond what data vendors typically cover,” NBIM said, adding that AI-generated summaries allow portfolio managers to respond faster to emerging risks.
Fund managers receive daily automated risk assessments covering newly acquired holdings, a system NBIM says has already helped it exit certain investments before wider market reactions triggered financial losses.
Chief Executive Nicolai Tangen described artificial intelligence as transforming how long-term institutional investors operate in increasingly complex global markets.
More Read
“Artificial intelligence is changing how we work as an investor,” Tangen said, noting that sustainability and governance considerations are now inseparable from financial performance.
The technology has proven particularly useful in analysing smaller firms in emerging markets, where corporate disclosures and international media coverage may be limited or published only in local languages.
NBIM’s adoption of AI comes as asset managers worldwide face mounting pressure from regulators and shareholders to demonstrate stronger ESG oversight while maintaining financial returns.
The fund reported annual profits equivalent to nearly $247 billion in 2025, buoyed by strong global equity performance. Nearly 40 percent of its investments are concentrated in U.S. stocks, including major holdings in technology giants such as Nvidia, Apple and Microsoft.
However, NBIM’s ethics-driven investment decisions have also sparked geopolitical debate. In 2025, the fund faced criticism from U.S. officials following its decision to divest from American machinery manufacturer Caterpillar and several Israeli banks over concerns linked to alleged human rights risks in Palestinian territories.
Norwegian authorities stressed that such divestments were based on financial risk assessments rather than political considerations.
More Read
Amid the controversy, Norway’s government introduced temporary guidelines limiting the fund’s authority to exclude companies from its portfolio while a broader review of its ethical investment framework is conducted.
Despite the policy review, NBIM said it would continue strengthening the integration between ownership responsibility and investment strategy, with artificial intelligence expected to play an expanding role.
Analysts say the move signals a broader transformation in global finance, where AI-driven analysis is increasingly becoming central to risk management, compliance and sustainable investing.
As sovereign wealth funds and institutional investors manage ever-larger and more complex portfolios, NBIM’s adoption of AI tools may set a precedent for how technology reshapes ethical investing standards worldwide.
