Investors urgently need to clarify what they believe about AI to make informed and consistent capital allocation and stewardship decisions, new research says.
Few investors have clarified what they believe about how AI should develop, even though six companies at the centre of the AI build-out make up about a fifth of the MSCI World index.
A new Silicon Boundaries research note sets out the key questions on which investors’ beliefs divide, with a free online tool that produces a draft statement of AI investment beliefs.
Investors’ values and beliefs about AI guide their stewardship, their voting, the mandates they give managers and what they exclude. When enough investors act on the same beliefs, they may also help shape how AI develops. But new research suggests that most investors have not yet set out clearly what they believe about AI. That needs to change, and quickly.
Of the world’s 200 largest pension funds and sovereign wealth funds, 99 have published investment beliefs, accounting for assets of about US$20 trillion. Of these, only one mentions AI, to say that it engages some of its portfolio companies to speed their adoption of it. Yet at the end of September Nvidia, Microsoft, Alphabet, Amazon, Meta and Broadcom, companies at the centre of the AI build-out, made up about a fifth of the MSCI World index.
To help investors rapidly develop and clarify their beliefs about AI, the Oxford Sustainable Finance Group at the University of Oxford and Theia Finance Labs have published a research note, “What do you believe about AI?”, and built an accompanying AI investor beliefs tool. The tool turns an investor’s answers to a series of questions into a draft statement of its AI investment beliefs.
Dr Ben Caldecott, Director of the Oxford Sustainable Finance Group and the Associate Professor of Sustainable Finance at the University of Oxford, said: “Clearly articulated beliefs let investors act consistently internally and in line with their clients and beneficiaries. They shape how investors vote on shareholder proposals, set mandates and exclusions, and decide whether to buy into new listings. Acted on by many investors, beliefs about what should happen also help decide what does happen. It is remarkable that so much capital has been invested in AI without most investors clarifying and then articulating what they believe about it.”
Dr Jakob Thomä, Co-Founder and CEO of Theia Finance Labs, said: “To help investors, we identify the key questions on which reasonable investors can disagree about AI. They include whether the most capable models should be open for anyone to run, who should decide when they are released, whether a model should do whatever its operator lawfully asks, and who should bear the cost when AI displaces human work. On questions such as these, the same company conduct can be a strength to one investor and a liability to another.”