SINGAPORE -- Technology companies, including social media giant Facebook and ride-hailing unicorn Grab, expressed support for a framework developed by Singapore aimed at strengthening oversight on the use of artificial intelligence.
The two companies are among more than 10 that publicly backed the framework after it was first published last year. A revised set of guidelines was released on Jan. 21 at the annual meeting of the World Economic Forum in Davos, in response to feedback from businesses.
"Singapore's Model AI Framework is a useful guide for organizations committed to AI governance and are looking to build foundational AI ethical principles into their product development and deployment processes," said Erin Egan, Facebook's vice president and chief privacy officer, in a statement.
Grab told the Nikkei Asian Review that it will use the document to educate and guide teams in the use of AI. "The model framework is a good start, especially in a region like Southeast Asia where AI technology is still nascent. We hope that other members of the global community can also contribute and provide their feedback to make this more comprehensive," a company spokesperson said.
Besides Facebook and Grab, other companies that have lent support to the framework include digital payments services Visa and Mastercard, financial institutions HSBC and DBS Bank, and U.S. drugmaker MSD, better known as Merck. Google has provided feedback but has not said whether it will support the document.
The framework spells out steps organizations should take to ensure responsible and ethical use of AI. It covers areas such as risk management, human intervention, interaction with stakeholders impacted by the technology and monitoring. It specifies that groups using AI should appoint a coordinating body or people to oversee the technology, and recommends setting up reporting systems to track its impact.
In businesses where safety is a critical issue, the framework proposes that organizations ensure humans can override AI systems where needed, and recommends regular reviews of potential harms from its use.
It proposes that groups using AI keep comprehensive records of the data generated to facilitate monitoring and audits, where needed. The framework also provides case studies of how companies have put principles of responsible AI use into practice.
"The AI model governance [framework] is not mandatory. It is a guideline. It is really [about] voluntary adoption, and I think many companies have seen it as very, very, valuable to actually adopt it as best practice in their way of delivering systems and processes using AI," said Tan Kiat How, chief executive of Singapore's Info-communications Media Development Authority.
Singapore is pushing for greater international oversight on the use of emerging technologies such as AI, as innovation threatens to outpace the ability of countries to set common standards governing their use internationally.
The U.S. earlier this month asked for feedback on a draft memorandum that it published on development of regulatory and nonregulatory approaches to sectors powered by AI. The memo covered themes similar to those of the Singapore framework, including risk-management, safety and security.
While the U.S. approach is meant to safeguard the country's national interests, Singapore has opted to gain support for its AI initiative by working with other governments and businesses simultaneously.
"This work, we believe, will pave the way for the next bound in global digital economy developments by fostering trust and strengthening collaboration between public and private sectors, and also with all other stakeholders," said Singapore's Communications and Information Minister S. Iswaran at the release of the country's updated AI governance framework.
Earlier on Jan. 21, Singapore, New Zealand and Chile concluded talks on a digital trade agreement. The pact establishs rules for e-commerce and new technologies and is aimed at promoting trade and investment. It includes a section on ethical use of AI governance frameworks, and advocates for transparent, fair and easy-to-understand use of AI. The deal parallels the principles laid out in the Singapore document.
Singapore officials hope more governments will sign on to the digital trade agreement, expanding its reach beyond the three signatories, and that more companies will back the proposals laid out in its AI governance framework.
Su Lian Jye, principal analyst at technology analysis specialist ABI Research, praised the city-state's approach in garnering support from countries and businesses for its AI governance agenda, but warned that it would take time to bear fruit.
Noting that AI powerhouses such as the U.S. and China are pursuing different paths for the technology, Su said a small country like Singapore may have difficulty helping others find common ground.
"Without help and influence from the private sector, the Singapore government will find it rather difficult to lead and shape global AI discourse," Su said.