SINGAPORE — With Artificial Intelligence (AI) set to change the way companies do business, Singapore has proposed a set of guidelines to encourage the ethical and responsible use of this technology.

The proposals, which are now open for pilot adoption and public feedback, seek to ensure two things: That decisions made by or with the assistance of AI are explainable, transparent and fair to consumers; and that their AI solutions are human-centric.

This Model Artificial Intelligence Governance Framework was announced by Communications and Information Minister S Iswaran at the World Economic Forum in Davos, Switzerland, on Wednesday (Jan 23).

What are human-centric AI solutions?

For instance, an online retail store wants operational efficiency. So, it turns to AI to fully automate the recommendation of food products to individuals based on their browsing behaviours and purchase history.

The guidelines say that while it does so, it also has to consider several factors. These include whether the product recommendations truly meet the needs of consumers and whether human intervention is needed in the event things go awry.

The store also has to regularly review its AI solutions to take into account changing societal norms. For example, given rising concerns over diabetes, the store might have to consider fine-tuning its AI solution to reduce the promotion of sugary drinks.

Why have such guidelines?

As part of efforts to become a digital economy and smart nation, Singapore is investing and building capabilities in AI. This includes the launch of a national programme called AI Singapore in May 2017 with a S$150 million funding to boost Singapore’s capabilities in AI.

As Mr Iswaran told the media in Davos: “It has the potential to affect many sectors in very profound ways. We want to make sure that Singapore is prepared for that in terms of our capabilities, technology investments, as well as our regulatory requirements.”

In a statement, the Info-communications Media Development Authority noted that human-centric AI solutions “enhance trust in and understanding of AI, as well as acceptance of how AI-related decisions are made for the benefit of users”.

Asked whether the set of guidelines might deter companies from setting up operations here, Mr Iswaran said that it is neither a “rule nor a regulation” and was borne out of a consultation with the industry.

As such, the guidelines are “well within the bounds of what they (companies) have already thought are important areas for them to focus on”, he said.

“On the contrary, I would say that we might even be able to come out as one of the jurisdictions with sound approaches to data management and the governance of artificial intelligence and other frontier technologies. In this case, more companies would want to be associated with us,” he added.