NY Common joins allocator push on company AI transparency

The $273 billion New York State Common has upped the pressure on portfolio companies to report on how artificial intelligence usage is contributing to layoffs, as AI governance becomes a growing focus in the proxy voting and engagement activities of asset owners.

New York State Comptroller Thomas DiNapoli, who is the sole trustee of the NYS Common, is demanding portfolio companies including Amazon, Salesforce, Meta and Pinterest – which have attributed job cuts to AI – to offer further information on the layoffs and any future workforce impacts.

In an opinion piece, he argued that it’s difficult for investors to assess whether changes related to AI are sustainable or beneficial for companies in the long-term without appropriate disclosures.

“[Companies] are eager to highlight productivity gains, but rarely provide transparency on AI’s consequences,” he wrote.

“Our state pension fund supports innovation that strengthens long-term growth and shareholder value, but durable growth requires more than technological ambition.”

NYS Common filed 28 shareholder proposals with portfolio companies in 2025, of which three are related to AI transparency.

“Investors need better information to judge whether AI-driven workforce changes reflect thoughtful strategy or short-term cost cutting that undermines long-term future performance,” he said.

DiNapoli joins a number of investors that have in recent years demanded more rigorous transparency practices around AI. CalPERS added a policy to its proxy voting guidelines this April that it may “withhold votes from director nominees where there is evidence of failed and/or insufficient oversight of AI-related risks”, while the $38 billion San Francisco Employees’ Retirement System updated its proxy voting principles to say that it would vote against directors if a lack of supervision on AI led to misuses that harm shareholder values. 

Railpen, the $45 billion UK pension for railway workers, also identified responsible AI as one of its stewardship priorities in a five-year plan to 2030. The fund is targeting companies in the EU, UK and US where it believes regulatory frameworks are creating the most risks and opportunities, and in sectors most prone to disruptions including IT, financial services and healthcare.

Norges Bank Investment Management, which is known for the prolific use of AI within its own organisation, published its house view on how portfolio companies could ensure responsible AI usage in 2023.

In a letter, chief executive Nicolai Tangen and chief governance and compliance officer Carine Smith Ihenacho wrote that companies need to be able to explain how the AI systems they use have been designed, trained and tested, and ensure that stakeholders can assess their risks.

The fund, which is the world’s largest asset owner, also highlighted AI’s long-term effects on the “human capital management” of companies as one development to watch.

The desire for more transparent AI practices in companies from allocators is translating to actions among asset managers, too. A Morningstar Sustainalytics report in 2025 found that the average support for the 15 AI-related shareholder resolutions it examined was 30 per cent, almost double the support (16 per cent) for the 400 resolutions on other environmental and social themes it sampled.

Investors are most commonly demanding information from companies on the way their board oversees AI usage, potential societal risks and their approaches to AI risk management.

If looking at significant resolutions specifically, the report observed that US asset managers’ support for those around AI (41 per cent) is higher than their support for significant resolutions around environmental and social issues as a broad category (35 per cent). With a pull-back on ESG sentiment in the country, the report suggested that US managers “see oversight of AI as a more financially material issue”.

European managers have a considerably higher support rate than US managers on AI-related significant resolutions (84 per cent).

The US managers also had significantly more dispersion, with some managers such as MFS, Fidelity, and Principal backing more than 70 per cent of the AI-related resolutions they voted on, while others like State Street and Vanguard didn’t back any of the resolutions.

Some common reasons for not backing AI-related resolutions include concerns that it would create unnecessary compliance burdens for the company or leak commercially sensitive information.