World Benchmarking Alliance raises the bar on human rights impact assessments and insufficient governance around the technology.
The World Benchmarking Alliance (WBA) has said it will update its Investor Statement on Ethical AI in February, in line with its commitment to drive the practical implementation of principles by companies.
In 2022, the WBA introduced the Collective Impact Coalition for Digital Inclusion (CIC) – an engagement campaign aiming to push for the adoption of ethical policies and practices around AI by tech companies.
WBA’s Digital Transformation Lead, Jan Rydzak, told ESG Investor that the updated statement on ethical AI would serve as an umbrella under which investors would be able to implement this collective push for more transparency, which underpins the whole CIC campaign.
“This new phase is expanding on the ask for companies to publish principles and demonstrate evidence of how they apply those principles”, he added. “This includes how they assess the impact of their AI systems, how the technology factors into their human rights assessments, and how they incorporate considerations around AI governance.”
The CIC campaign builds on the WBA’s Digital Inclusion Benchmark, which assesses 200 firms on the extent to which they enable access to digital technologies. As of September 2023, 52 of the 200 companies (26%) had adopted ethical AI principles, marking an increase from 33 the previous year.
Despite this progress, the alliance said its benchmark results showed that the pace of progress on AI governance remained insufficient. It had previously noted that companies’ oversight mechanisms around AI often remained poorly explained and understood.
This, Rydzak explained, is one of the reasons why the WBA decided to expand the CIC.
“In short, we’re hoping the update will raise the bar for all companies [and] create a new baseline they will strive for beyond high-level principles,” he added.
“We hope this interaction will stimulate the race to the top on AI. If a single company shows it can be a pioneer and disclose a larger amount of critical information, it can become the reference point for others to do the same.”
Human rights laggards
The WBA’s original investor statement had stressed the need for investee companies to implement policies and mechanisms to ensure the ethical development and application of AI, guided by respect for human rights.
Yet, a recent round of research conducted by the alliance found that out of 52 companies with publicly disclosed AI principles, only six were able to show that they factored the impacts of the technology into their human rights assessments, which Rydzak said was a “disappointing finding”.
“AI is such a ubiquitous technology, it doesn’t stand to reason that companies can remain silent on what due diligence they exercise,” he added. “There is a considerable risk of fragmentation on AI. The combination of strong interest and uncertainty in the regulatory space means that a lot of investors are not quite sure what to ask of companies.”
The WBA’s updated statement will also aim to rally larger groups of allies and investors to the CIC, as well as civil society members. Upon the CIC’s launch, the alliance had brought together a group of 33 investors – including CIC initiative investor co-leads Fidelity International and Boston Common Asset Management.
Other participants from the investor community include Aviva, Cardano, HSBC Asset Management, Robeco and Schroders. The participants are responsible for more than US$6.9 trillion in assets under management.
The WBA will make frequent updates to the data, which will enable both the investors and the public at large to track companies progress on ethical AI practices.
“We’ll be publishing data updates every six months as opposed to annually, and will also aim to publish full progress reports once a year,” Rydzak said. “It’s important for there be an up-to-date view on this.”
This week, the World Economic Forum’s AI Governance Alliance stressed the need for a global effort to create equitable access to AI, publishing three new forum papers with recommendations on building safe systems and technologies, ensuring responsible applications and transformation, and advancing resilient governance and regulation.