Baidu becomes first Chinese firm to join US-led safe AI consortium

Baidu becomes first Chinese firm to join US-led safe AI consortium

Internet search giant Baidu has become the first Chinese company to join a US-led consortium dedicated to the safe development of artificial intelligence.

The Partnership on AI to Benefit People and Society (generally known as the Partnership on AI) was established in 2016 to “realise the promises, minimise the challenges, and better understand the impacts of artificial intelligence”.

Its 70+ members include tech providers such as Amazon, Apple, DeepMind, Facebook, Google, IBM, Intel, Microsoft, Nvidia, PayPal, Salesforce, and Software AG; consultancies and service companies, such as McKinsey and Accenture; academic institutions, including MIT and the Future of Humanity Institute at Oxford University; and a variety of other organisations, including Amnesty International and Human Rights Watch.

With China often associated with issues such as internet censorship, citizen surveillance, and compulsory social ratings, Baidu’s commitment to the organisation is both intriguing and a sign that the Partnership is becoming a more global, rather than US-centric, force.

According to the Partnership, Baidu has joined because it believes that it is necessary to “take relevant measures to ensure the safety and controllability of the technology”.

According to the Partnership, Baidu has kicked off series of internal initiatives aimed at limiting AI’s unintended consequences.

For example, ‘safety and security’ is Baidu’s first principle in the development of autonomous driving, and recommended best practices for production-scale autonomous driving systems were published in the industry’s first Apollo Pilot Safety Report in July 2018.

Apollo, which is China’s largest open-source autonomous driving platform, is one of Baidu’s three major AI offerings, along with DuerOS, its voice-enabled digital assistant, and Baidu ABC, a smart cloud for the enterprise market.

Speaking out

Ya-Qin Zhang, president of Baidu Inc, said, “As AI technology keeps advancing and the application of AI expands, we recognise the importance of joining the global discussion around the future of AI.

“Ensuring AI’s safety, fairness, and transparency should not be an afterthought, but rather highly considered at the onset of every project or system we build.

“The impact of a transformative technology like AI goes beyond borders, so we are looking forward to both sharing our own insights and learning from our international peers.”

Baidu’s admission to the group represents the beginning of the Partnership’s expansion into China – according to the consortium, which said, “We will continue to add new members in China and around the world as we grow, bringing together for-profit companies and civil society organisations.

“Baidu’s AI leaders will join the Partnership’s existing Working Groups and forthcoming lines of programming, contributing to research projects that are intended to develop the policies, tools, and principles that will inform and drive responsible AI development and deployment.”

The Partnership has just launched three new Working Groups: AI, Labour, and the Economy; Safety-Critical AI; and Fair, Transparent, and Accountable AI.

Terah Lyons, executive director of the Partnership on AI, said, “Admitting our first Chinese member is an important step toward building a truly global partnership. The growth and scope of work on AI in China is extensive, and any conversation about the future of AI that does not involve China is an incomplete conversation.

“I look forward to seeing our members work together across borders, sectors, and disciplines to help shape the future of this critical technology, enabling the best possible outcomes for society.”

Internet of Business says

Despite the deepening trade war, which may soon affect all Chinese exports to the US, partnership with China is critical to many American companies, not least because of the vast manufacturing base it offers to the technology sector.

AI and autonomous driving are a big part of that relationship. For example, Waymo – the driverless car division of Google parent Alphabet – confirmed in August that it had decided to join Tesla by opening operations in China, which is now the world’s biggest automotive market.

Alphabet has registered a company in Shanghai called Huimo Business Consulting. As well as being a near homophone for Waymo, ‘Huimò’ is Mandarin for ‘emblem’.

The name is certainly emblematic of change. According to the new company filing, potential operations for Huimo include designing and testing autonomous vehicles parts and products in China, as well as offering consultancy services – as the name suggests.

In July, Intel’s computer vision subsidiary Mobileye joined the Apollo ecosystem, in a move that saw Baidu and its partners adopt Mobileye’s Responsibility, Sensitive Safety (RSS) model into their autonomous vehicle programmes.

That said, AI remains a controversial topic with regard to China. With 1.4 billion people, few data privacy protections, and, from 2020, a nationwide compulsory social ratings scheme that seeks to punish non-conformity, China represents a testbed for the technology rivalled only by Facebook in terms of numbers.

Last week in Shanghai, for example, hardware giant Huawei launched a range of new AI and IoT products at its Connect 2018 event, including dedicated AI chips, mobile data centres for autonomous cars, and more. It also revealed that it has been selling AI-enabled, software-controlled cameras and other smart, connected technologies to some of the most repressive regimes in the world.

Judging by some of the presentations at the event – attended by Internet of Business editor, Chris Middleton – Huawei sees AI in smart cities as being about policing and security at least as much as it is about sustainability and green technologies.

China itself is a vast market for US technology, which has left companies such as Google on the horns of an ethical dilemma: should they pursue censored search technology against the wishes of their own employees? Or stay out of a market in which more than 700 million people are online, with the vast majority using apps such as WeChat to communicate and buy goods?

A number of US platforms, including Facebook and Google, are blocked in China, while other services, such as Microsoft’s Bing search engine, are available in local forms.

Evidence emerged this week that Google is attempting to justify re-entering China on ethical grounds – reversing its previous stance.

When news of Project Dragonfly, Google’s experimental censored search technology for the Chinese market, leaked in August, the company was slammed for abandoning its policy of walking away from China over state censorship. However, on 15 October, Google CEO Sundar Pichai claimed that Dragonfly could serve “well over 99 percent of queries” and help Chinese citizens find essential information, such as treatments for cancer.

Pichai will have to work hard to convince his own staff that this wasn’t the most brazen of about-turns and pure economic opportunism.

Chris Middleton
Chris Middleton is former editor of Internet of Business, and now a key contributor to the title. He specialises in robotics, AI, the IoT, blockchain, and technology strategy. He is also former editor of Computing, Computer Business Review, and Professional Outsourcing, among others, and is a contributing editor to Diginomica, Computing, and Hack & Craft News. Over the years, he has also written for Computer Weekly, The Guardian, The Times, PC World, I-CIO, V3, The Inquirer, and Blockchain News, among many others. He is an acknowledged robotics expert who has appeared on BBC TV and radio, ITN, and Talk Radio, and is probably the only tech journalist in the UK to own a number of humanoid robots, which he hires out to events, exhibitions, universities, and schools. Chris has also chaired conferences on robotics, AI, IoT investment, digital marketing, blockchain, and space technologies, and has spoken at numerous other events.