Opening Remarks by Ms Beate Trankmann at the International AI Cooperation and Governance Forum 2022
December 9, 2022
尊敬的王希勤校长 (WANG Xiqin, President of Tsinghua University)
尊敬的李萌副部长, (LI Meng, Vice Minister for Science and Technology)
徐晓兰副部长, (XU Xiaolan, Vice Minister of the Ministry of Industry and Information Technology)
马升琨副司长, (MA Shengkun, deputy director-general of the Department of Arms Control of the Ministry of Foreign Affairs)
Siddarth Chatterjie, UN Resident Coordinator in China
Shahbaz Khan, Director of UNESCO Beijing and Representative to China,
Ladies and Gentlemen,
On behalf of the United Nations Development Programme in China (UNDP) I am delighted to join this third annual International AI Cooperation and Governance forum.
Digitalization is transforming our world. With new technologies including AI, rapidly reshaping our economies, societies and how we live, development has to be reimagined for this digital age, to effectively address the many urgent challenges we face.
Indeed, in 2022, those challenges worsened. From devastating floods in Pakistan and the Americas, to deadly heatwaves in China, climate disasters are growing more frequent and extreme. Global hunger has also hit new highs, with 276 million people facing severe food insecurity, doubling from two years ago. Food, along with energy prices are soaring, squeezing those most vulnerable, especially during winter.
Our world is in crisis. So - we need all the help we can get to realise the Sustainable Development Goals (SDGs) to end poverty and ensure our planet can continue sustaining us all, by 2030.
Recognizing this, UNDP’s Strategic Plan for 2022 through 2025, highlights digitalization as a “key enabler” to maximize and scale development impact, and outlines our engagement in this area. If governed wisely, emerging technologies have the potential to provide vital, real-time information and forecasts that can accelerate solutions to put the SDGs back on track.
Already, we can see how AI is being leveraged to help boost renewable energy, through the use of weather forecast data to predict wind and solar power production, allowing their electricity deliveries to be scheduled and used more efficiently.
AI also gathers real-time data to lift crop yields, save water, cut food production costs and combat pests. Using drone-generated data, for instance, the UN and others are working to improve pest management in agriculture.
With so much potential, demand for AI has surged recently. The Economist reports job ads by S&P 500 companies for AI and machine learning experts were about 10 times higher in the last three months than the first quarter of 2020. This rapid rise in popularity makes it even more vital that AI is carefully managed, to help solve, rather than worsen, our problems.
While AI should not be over-regulated, to avoid stifling innovation, we must take care to protect society against potential abuse.
An adequate governance framework for AI as well as oversight and ethical guidance are key, not only in privacy protection during data collection, but also to ensure that AI does not deepen societal divides and discrimination. Because the algorithms behind it risk carrying the same biases as the people programming them.
By way of example, let me highlight two areas where we can see these risks emerging:
The first is AI’s impact on the future of work – not only in automation, but also AI based decision-making in recruitment, which could expand inequality. In 2018, Amazon reportedly abandoned an internal AI recruitment program after finding it rated male candidates higher than women for technical positions, because men had more frequently been hired for these jobs previously, and the algorithm worked to reinforce this pattern.
The second area is gender equality – a topic I am particularly passionate about. In addition to bias in employment, which I just outlined, AI also excludes women in other ways, from finance, to health. In 2021, Women’s World Banking found AI credit rating systems used by financial service providers globally are likely to discriminate against women, which risks widening the $17 billion gender credit gap. Even Apple co-founder, Steve Wozniak, reported receiving a credit limit ten times higher than his wife from Apple Card, despite sharing the same bank account. Meanwhile, in health, it was found that a lack of female patient data in training computer-aided diagnoses (CAD) systems may lead to lower accuracy in diagnosing female patients, particularly given men and women often show different symptoms for the same diseases.
These two areas will be explored in this year’s forum through two sessions anchored by UNDP. I look forward to gaining further insights during these, as well as other, deep-dive discussions and discovering new ways of addressing relevant challenges with all of you.
We must come together – including the tech sector, government and media – to develop a global framework and standards to effectively govern AI, with multilateral cooperation. AI experts must try to explain AI, so it can be better understood – including the human biases it reflects. This would allow more effective regulation by governments, with safeguards to respect and protect everyone. The media can also play an important watchdog role, raising awareness about individual rights and risks regarding AI.
In closing, let me express my sincere thanks to Tsinghua University’s Institute for AI International Governance for organizing this important forum, and to all the supporting partners involved.
While AI can play a powerful part in sustainable development, it can also do the reverse. We must be proactive to unlock its potential, but also proceed with caution and cooperation, so no one is left behind.
祝本次论坛圆满成功 (I wish the forum a complete success)!
Remarks at the Thematic Forum on Consultation and Global Governance, organized by the Chinese People's Political Consultative Conference
Remarks by Ms Beate Trankmann on Using Data to Strengthen Governance Systems at UN World Data Forum
Tsinghua AI Forum: Multilateral and Multistakeholder Global Platform Needed to Build a Balanced and Inclusive AI Governance Framework
A multilateral and multistakeholder global platform is needed to build a balanced and inclusive AI governance system, agreed world-leading experts at the 2021 Tsi...