Google CEO Sundar Pichai calls for 'proportionate' regulation of artificial intelligence

There is no question in my mind that artificial intelligence needs to be regulated - Sundar Pichai

Google CEO Sundar Pichai has called for ‘proportionate' regulation of artificial intelligence to combat the "negative consequences of AI" and "nefarious uses of facial recognition".

"There is no question in my mind that artificial intelligence needs to be regulated. It is too important not to. The only question is how to approach it," Pichai wrote.

He continued: "Good regulatory frameworks will consider safety, explainability, fairness and accountability to ensure we develop the right tools in the right ways. Sensible regulation must also take a proportionate approach, balancing potential harms, especially in high-risk areas, with social opportunities."

Are you working on exciting AI or machine learning projects in your organisation? Enter them - and yourself and colleagues - in Computing's upcoming AI & Machine Learning Awards 2020. Now open for entries!

AI will become an increasingly deeply embedded element of people's everyday lives, he added, with AI used across the medical profession to monitor health conditions and, at some point in the future, self-driving vehicles potentially making life or death decisions. These will require "rules that consider all relevant costs and benefits", he suggested.

Those rules, he hinted, could be based on Google's own AI Principles, introduced in 2018. That followed criticism over the company's involvement in military work on drone footage image recognition. However, Google's attempts to put together an ethics board to govern its AI-related activities have proved challenging.

Pichai made his pitch in a Financial Times opinion article today, which he followed up in a speech in Brussels today. "There's no question in my mind that artificial intelligence needs to be regulated [but] we don't have to start from scratch," he said today.

Pichai's call comes as the European Union prepares to formally release its white paper on AI, after a December draft was leaked last week.

That contained the proposal for a five-year ban on the use of public facial recognition technologies - among five potential regulatory options. That would lend time in which a "sound methodology for assessing the impacts of [AI and image recognition] and possible risk management measures could be identified and developed".

There would be some exceptions for, for example, research and development purposes. Furthermore, any European regulation would be unlikely to prevent the use of images scraped from social media accounts of Europeans and added to facial recognition databases elsewhere, for whatever purposes.

The EU will set out its plan for legislation on AI in mid-February.

The facial recognition ban proposal comes at the same time that it was revealed that the New York Police Department (NYPD) in the US has been using a tool to match photos of people with images published online.

According to the New York Times, Clearview, a Peter Thiel-backed start-up has provided the tool to enable NYPD officers to identify almost anyone.

"You take a picture of a person, upload it and get to see public photos of that person, along with links to where those photos appeared.

"The system - whose backbone is a database of more than three billion images that Clearview claims to have scraped from Facebook, YouTube, Venmo and millions of other websites - goes far beyond anything ever constructed by the United States government or Silicon Valley giants," the New York Times claims.

The application has already been used to identify people involved in shoplifting, ID theft, credit card fraud and even murder, it added, with more than 600 police officers now using the tool.