Investors Need Better Information About How Companies are Using AI 

The IPOs of companies like SpaceX, OpenAI, and Anthropic promise to upend financial markets and generate eye-popping returns for many investors. However, the volatility and uncertainty surrounding these companies underscore another challenge: investors are struggling to value the AI transformation, and to account for how its impacts, risks, and opportunities will appear in a company’s bottom line.

As AI is transforming industries, from manufacturing to services, investors are looking beyond these major cases to assess how it will reshape business and society, what the major risks are, and how they should account for these unprecedented changes in portfolio management. But the effects of AI are complicated and difficult to predict. They also vary widely across sectors, companies, and regions.

According to the 2026 Stanford University Human-Centered AI Index Report, global corporate investment in AI more than doubled in 2025, with productivity gains ranging from 15% in customer support to 50% in marketing output. At the same time, the risks are broadening, from workforce disruption to environmental footprint, along with an ever rising count of documented AI incidents.

In conditions of greater certainty, investors and providers of capital can evaluate a company’s prospects and price its value using financial filings and independent research. But to evaluate how companies are responding to complex, far-reaching developments such as AI, they need information beyond what’s included in traditional financial filings.

Investors need to understand how risks and opportunities show up in a company’s operations, products, and value chain. This includes risks related to nebulous issues like public accountability, fast-moving regulation, liability, and workforce transformation. Investors are looking to evaluate how these issues are overseen and managed in a company’s governance structure. They also need to understand a company’s overall strategy, where it sees challenges or competitive advantages, and how its approach compares to competitors.

Having this information will help investors price in risks and opportunities that don’t appear on a company’s balance sheet. And it can give them better insight into which companies have the leadership, strategy, and guardrails to navigate the uncertainty of AI responsibly—to generate financial returns without potentially catastrophic organizational or societal disruptions.

Better AI disclosures also benefit other stakeholders. Governments, insurers, civil society organizations, and safety researchers all need high-quality information from companies to make informed decisions. Some of this information is already included in financial and sustainability reports, but much deeper insights are needed than are currently disclosed.

Reporting Standards

Over the past 25 years, several interoperable standards have emerged to define how companies should report on this broader set of impacts, risks, and opportunities. Most prominent are the SASB Standards, the ISSB Standards (IFRS S1 and S2), the European Sustainability Reporting Standards (ESRS), and the Global Reporting Initiative (GRI) Standards.

These standards are increasingly mandated by law, integrated into stock exchange listing requirements, or accepted as best practice. According to KPMG, 96% of the world’s largest 250 companies report on their sustainability impacts, with SASB standards particularly dominant among companies in the Americas, and a growing number of companies preparing to adopt the ESRS.

Companies do not need a new reporting standard to improve their AI-related disclosures. They do need enhanced guidance on how to apply existing reporting standards in the context of AI.

Draft Disclosure Recommendations

Today, PAI is publishing a draft of Disclosure Recommendations to help companies identify material information about the impacts, risks, and opportunities of AI to include in their financial and public reporting. These Disclosure Recommendations are:

READ THE RECOMMENDATIONS

• Industry-agnostic, but designed to capture nuances from domains in which AI is deployed, such as healthcare, financial services, retail, infrastructure, energy, social media, and the public sector.
• Intended to improve the quality of corporate-level information that informs decision-making, such as capital allocation decisions.
• Structured to support, rather than add to, the requirements of existing reporting standards.
• Designed to be used alongside the Corporate AI Risk Assessment Framework that we published earlier this year. That framework defines what good corporate governance and management look like, while these recommendations help companies report externally.

Investor-focused reporting is one element in the larger field of AI transparency. System cards, regulator-required reports, and incident reports are other tools, aimed at different audiences, that all support the goal of increased trust and transparency.

To develop these recommendations, PAI consulted with a mix of investors, companies, civil society organizations, and other experts. We plan to refine them over time based on feedback and as reporting practices evolve.

SHARE FEEDBACK

We welcome feedback on these draft Disclosure Recommendations and the opportunity to work with companies, investors, and other stakeholders interested in testing them. For more, please get in touch with Sam Wallace.