The world's largest sovereign wealth fund uses Claude to "minefield": problem companies revealed within 24 hours

The world’s largest sovereign wealth fund—the Norwegian Government Pension Fund Global (NBIM)—says the organization has been using Anthropic’s Claude model to screen for ethical issues in its investments.

The Norwegian Government Pension Fund Global (officially called Norges Bank Investment Management, NBIM) is operated by the Norwegian Central Bank and currently has a total size of about $2.2 trillion. By 2025, its investment returns are expected to reach approximately $246.9 billion.

The fund has invested in over 7,200 companies worldwide, holding about 1.5% of the world’s publicly listed stocks, giving it significant influence in global markets and ESG (Environmental, Social, and Governance) investing. The fund leverages its influence and voting rights to set expectations for the companies and markets it invests in.

In its annual responsible investment report, the fund management team stated that they are now using artificial intelligence (AI) to provide insights on governance and sustainability to portfolio managers, helping to identify potential reputational and ethical risks.

NBIM says this technology broadens the scope and scale of information they can analyze, enabling them to “identify major risks more quickly.”

A spokesperson for NBIM said that their ESG risk monitoring team began using Anthropic’s Claude model in their daily work before November 2024, and that the tool has become “an important assistant in monitoring ESG risks across our entire investment portfolio.”

In the report released on Thursday, NBIM also announced that in 2025, they deployed large language models to conduct comprehensive screening of all new companies entering their stock portfolios on their “first day.”

The report states: “These tools help us quickly scan large amounts of public information, far beyond what traditional data providers offer. When risks emerge around key topics, the large language models conduct deeper searches and provide summaries with background information.”

NBIM receives AI-generated daily investment risk assessments from the previous day, allowing the team to consider risk mitigation immediately. “Within 24 hours of investment, the AI tools flag companies that may be involved in issues like forced labor, corruption, or fraud.”

“Many times, this information has not yet been captured by international media or data providers. We conduct manual reviews before making investment or risk decisions. In several cases, we sold these investments early, before the market widely reacted, avoiding potential losses.”

NBIM states that this AI approach is especially valuable for researching small and medium-sized enterprises in emerging markets, where relevant news often only appears in local media and in local languages.

Nicolai Tangen, CEO of NBIM, said in the report, “AI is changing the way we work as investors. Sustainability and corporate governance are closely linked to financial performance,” and pointed out that “the world will remain complex and uncertain.”