A vision for AI: Innovative, Trusted and Inclusive

By Mario Romao and Derek Waxman, AI Policy

Da Jia Hao! – Good morning to all – it was with these words that Abigail Wen from Intel’s Artificial Intelligence Products Group opened her presentation “Top AI breakthroughs you need to know about” at the Artificial Intelligence Conference, co-presented by O’Reilly Media and Intel Corporation held from June 18-21, 2019 in Beijing. The conference, under the banner “Break Barriers between Model and Reality” brought together global AI experts to bridge the gap between AI developments in research and their commercial applications in business and industry.

But what this blog is really about, is how Abigail ended the session: presenting Intel’s policy vision for the future of AI. Intel believes that the future of AI should revolve around three pillars:

  • Innovative — meaning that the future of AI requires constant innovation;
  • Trusted — meaning that the future of AI requires commitments to privacy, security, transparency and explainability;
  • Inclusive – meaning that the future of AI requires using ethical design and addressing social impact.

With AI, governments have the opportunity to address large societal challenges, while spurring economic growth, addressing citizen’s concerns and identifying needs for regulatory intervention. By focusing on Innovation, Trust and Inclusiveness, Intel invites governments to:

  • Commit resources to the development, adoption and implementation in AI and foster international collaboration between industry, academia, government and associations;
  • Remove barriers to the responsible access to data which will help machine learning and deep learning reach their full potential. Moreover, as autonomous decisions made by algorithms may affect individuals and their private lives, more automation should not translate to less protection;
  • Mitigate unintended consequences and ride on the potential of AI, as the social implications of computing have grown and will continue to expand while more people have access to implementations of AI.

At the margins of the conference, several members of Intel’s Government Markets and Trade Group – Terry Zhang, Ning Zou, Derek Waxman and Mario Romao – had the opportunity to discuss AI and data protection policies. We also exchanged views on Intel’s commitment to advancing uses of AI that most positively impact global society (e.g. healthcare) as well as the need to harness the power of data while also addressing a wide range of data protection concerns. We highlighted the role of privacy preserving technologies and federated machine learning to optimize data openness and availability as well as privacy protections. We also touched on the latest developments on Ethics and AI, namely the recent “Ethics Guidelines for Trustworthy AI” from a group of European experts under the auspices of the European Commission and the “Governance Principles for ‘Responsible AI” prepared by Chinese experts under the auspices of the Chinese Ministry of Science and Technology.

The O’Reilly AI Conference and adjacent discussions with Chinese stakeholders was a tremendous success, enabling dialogue and debate on how government and industry can move forward with the promise of AI while investing in key technologies and policies to protect data and personal privacy. Beijing was the third stop of Intel’s AI Policy World Tour (see what happened in Berlin and Ottawa) an initiative from Intel which aims at stimulating AI policy discussions across several regions of the world.