On 9 October 2020, the Center organized and co-hosted a webinar with the AI Policy Initiative (Seoul National University), and the Institute for Future Initiatives (The University of Tokyo) on the topic of “Data, AI Governance, and COVID-19: Medium and Long-Term Perspectives for Asia”. The webinar built on contributions from an earlier joint report published in September 2020 that mapped out China, South Korea, Singapore, and Japan responses to COVID-19 and its potential challenges from an AI, data governance and human rights perspective (https://ifi.u-tokyo.ac.jp/…/uploa…/2020/09/report_200916.pdf).
The webinar started with Prof HakSoo Ko (Co-director, SNU AI Policy Initiative) giving the opening address. This was followed by the session 1 "Lessons learned for AI and Data Governance" moderated by Prof Yong Lim (SNU), panelists included our centre director Prof Mark Findlay, Prof Arisa Ema (University of Tokyo), and Prof Yi Zeng (Chinese Academy of Science, BAAI). The second session on "Changing Lifestyles & Governance" was moderated by Prof Arisa Ema, and Prof Sangchul Park joined the panel. The webinar concluded with Prof Hideaki Shiroyama (Vice Director, Tokyo University Institute for Future Initiatives) delivering the closing remarks.
The panelists shared with participants their further reflections on the employment of these COVID measures, and what this emerging trend of technology and big databases signal for the future of AI governance and ethics. They covered a broad range of topics in the two sessions and weighed in on topical issues concerning #DataPrivacy, #StateSurveillence, #ContactTracingApplications, and #GoodGovernance. Panel members also engaged in a lively discussion on the durability of these COVID measures and highlighted several factors that may run counter to their effective termination in the foreseeable future.
We thank all the panel members and audience for the meaningful dialogue session.
Opening and Welcoming remarks
Professor Haskoo Ko, Seoul National University
Despite the difficult prevailing circumstances, the webinar was opened on a positive note by Professor Haskoo Ko who applauded how institutions have managed to adapt to living with COVID-19 through innovative switches to webinar events. Professor Ko communicated that the presence of today’s panellists, and the high attendance rate of participants logging in from various countries showed how having productive dialogues were still possible – he further emphasised that the high turn up rate of this event would not have been possible if the event was hosted offline.
Professor Ko went on to reflect how Asia has been handling the virus relatively well in terms of safeguarding its general population health and its economy. However, he expressed that since Asia is made up of different countries across its continent, it is important to acknowledge the different phenomenon and policies taken across the different states. Professor Ko commented that this webinar aims to give valuable insights as to what has transpired in Japan, China, South Korea, and Singapore and what we can and should do to deal with the pandemic in more effective ways.
Session I: Life under COVID-19: Lessons Learned for A.I. & Data Governance
Moderator: Yong Lim (Seoul National University)
Panellists:
Mark Findlay (Singapore Management University)
Yi Zeng (Chinese Academy of Science)
Sangchul Park (Seoul National University)
Question 1: Responses to COVID-19
Professor Yong asked each panellist to name one positive lesson that has emerged out of their jurisdiction in relation to the pandemic, and how their respective government has responded to the public health crisis.
Professor Yi Zheng kickstarted the discussion by explaining how the pandemic has brought forward an opportunity for scholars and the Chinese government to interact with each other in a positive direction. He communicated how everyone involved is learning from each other in the midst of adversity, and are collaboratively seeking out solutions to deal with the crisis for humanity’s overall benefit. Professor Yi further highlighted the role of society and how public feedback have been used constructively to produce better coordination and regulation outcomes – this has achieved greater protection and safeguards for public health. Finally, Professor Yi emphasised that everyone needs to find their own role in this symbiotic society, and that every individual should take responsibility for this multi-stakeholder framework powered by information technology.
In recollecting the positive lessons learned from fighting COVID-19, Professor Mark Findlay reflected that Singapore has had considerable success in the virus containment aspect. Professor Findlay highlighted the low mortality rate in the country and attributed it to the government’s integrated policy approach. He further considered that Singapore has fared well from the perspective of promoting good governance, and remarkably, the state has acknowledged its critical role and the role of the private sector in producing effective pandemic intervention methods. Professor Findlay then turned to emphasize that the state should also dedicate resources to ensuring that citizens’ trust in the government is maintained. He cautioned that if citizens are confused or misunderstood as to the impact of these COVID control strategies, trust would be damaged and the measures adopted would be less successful.
Prior to the pandemic, Professor Arissa Ema recollected that the Japanese government had proposed an initiative that pushes for the founding of a smart society. This concept was identified as Society 5.0 and it is defined as a human-centric society that focuses on economic development and the resolution of social issues through a system that integrates both cyber and physical space. Professor Ema raised that although this concept faces increasing challenges today, the pandemic had also ironically promoted its fulfilment in a separate sense. For example, concerning telemedicine, COVID 19 has accelerated the approval and use of remote diagnosis in Japan. Professor Ema added that the onslaught of COVID 19 has also urged all relevant stakeholders to build positive government, industry, and academic networks. In addition, the public health crisis highlighted the important role of the state in securing principles of privacy, security, fairness, transparency, and accountability.
Professor Yong Lim drew the discussion to a close by drawing recognition to the fact that South Korea has responded well to the pandemic by amending existing MERS legislations that were already in place. This has allowed the South Korean authorities to coordinate timely response to the outbreak and it has prevented the country from using ad-hoc, super-legal emergency powers – reliance on this legislation served as an effective starting point for the rolling out of other necessary amendments.
Question 2: unwillingness of Japanese citizens to sacrifice their human rights (e.g., privacy rights) stemmed
The next question was posed to Professor Ema from Japan. Professor Yong Lim queried whether the unwillingness of Japanese citizens to sacrifice their human rights (e.g., privacy rights) stemmed from a public trust issue with the Japanese authorities or a fear of stigma and ostracization in Japan?
Professor Ema first explained that despite the lack of legal enforcement, Japanese citizens were quick to exercise great self-restraint – as an example, she cited the voluntary closure of schools and other facilities in Japan although such closures were not legally mandated/enforced. However, she reflected and contrasted citizens’ voluntary attitudes against the low uptake rate of Japan’s contact tracing application, COCOA. She lamented that although the application is privacy-focused and anonymous in its design, Japanese citizens remained hesitant to download the application on their phones. Professor Ema acknowledged that this could potentially be attributed to the lack of trust in the state authorities and their desire to protect their privacy rights. However, she also believes that an element of peer pressure exists in the Japanese culture that may have influenced on the poor uptake of the application. She communicated that Japanese citizens may feel compelled to hide their infection from the public knowledge for fear of repercussions to themselves. Professor Ema then pointed out that it would be interesting to learn whether this culture is similarly prevalent in the other countries where our other panel members are located.
Professor Yong agreed with Professor Ema that the cultural aspect of AI is an important discussion topic to get into.
Question 3: AI as a tool to identify and minimise vulnerability risks
The next question was directed at Professor Findlay. Professor Yong asked Professor Findlay to elaborate further on a stated point in his contribution that we should be cautious that AI is not used to exacerbate existing structural problems but utilised to identify and minimise vulnerability risk. Professor Yong wishes to know how such outcomes can be achieved and how this can be translated into concrete policies.
To illustrate his point on minimizing risk posed to vulnerable groups, Professor Findlay cited the outbreak that occurred in Singapore’s migrant workers’ dormitories. He stressed that out of a total of 56,000 reported infections in the country, 55,000 of those infected were migrant labourers. Professor Findlay went on to explain how migrant labourers in Singapore are a socially anonymous group who live in conditions/residential environments that prevent the effective implementation of control practices like social distancing. Professor Findlay raised that this is a clear example of when diagnostic prediction had failed to foresee the risks posed to these discriminated individuals. This lack of diagnostic prediction also resulted in the untimely deaths of elderly persons residing in institutional care homes in the UK, and high infection rates of prisoners incarcerated in the US. Professor Findlay is of the view that if sensible diagnostic risk profiles were developed at an earlier point, authorities would have been able to point out the susceptible cases and implement more appropriate strategies accordingly – pandemic-related deaths and infection rates would have been greatly reduced as a consequence.
Professor Findlay then turned to explain that there are also cultural foundations in the way Singapore citizens identify with their personal data and corresponding rights. He explained that there are two existing structures and strategies of surveillance adopted by Singaporean authorities: Firstly, track and tracing, and secondly, safe entry QR codes. He went on to express that the uptake of the TraceTogether application is relatively low as citizens are concerned about the storage and use of their personal data. However, Safe Entry is more widely accepted in the country and its popularity may be tied to the concept of individual self-determination. Ultimately, people have different perceptions on these technologies and these perceptions are deeply rooted in culture.
Question 4: the temporariness of COVID-19 measures
The next question posted by the moderator relates to the durability and “temporariness” of these adopted measures. Professor Yong Lim explained that when these measures were initially introduced to the public, it was implied that the strategies were only temporary in nature. However, a creeping realisation has dawned upon us in the recent months that the virus may live with us for a longer time than expected. The question that Professor Yong Lim posed to the panellists then is what should be done in relation to the measures that are currently in operation – Should we conduct a post-crisis review and embark on the phasing out of certain measures, or retain these measures since some of them have proven effective in containing the virus? Panel members are also asked to consider what should fall to be the reasonable criteria to determine whether a policy should be kept in place or terminated.
Professor Yi Zheng explained that people in China define themselves by the connections they share with others in the community. As such, Chinese citizens’ reaction to the state’s pandemic containment measures had also informed and influenced the government's overall regulatory approach.
Despite the dwindling infection rates of COVID cases in China, Professor Yi Zheng stressed that it would be difficult to retreat some of these surveillance measures. He highlighted that citizens themselves are hesitant to withdraw the use of these technologies as many expressed their concerns that it is still unsafe to do so. Additionally, it would be difficult to roll-back specific technologies as many of these tech are used in combination with others (e.g., facial recognition and temperature infrared sensors). Professor Yi also highlighted that although face masks are not legally mandated in China, many citizens view mask-wearing as a way to protect themselves and contribute to the protection of other members of society – the pandemic has altered society in relation to how individual citizens view their relationship(s) with each other and citizens’ self-reinforcement of these safety measures may ultimately end up stricter than the government's own regulations.
Professor Ema responded that Japanese citizens share the same sentiments as the Chinese. In Japan, huge importance is placed on protecting oneself and one’s community from harm’s way. Individuals are also mindful and reluctant to inconvenience others and as such, citizens fear passing on the virus to others in their community. A culture of “peer pressure” is also prevalent in the state – infected individuals face discrimination from wider society, and are reproached and shamed by others for spreading the virus. Many Japanese citizens also reported feeling more afraid of infecting others with COVID as opposed to catching the virus themselves.
Professor Yong agreed with the above panellists responses – he recollected the emphasis placed on “culture” in their above points and agreed that culture will play an important role in the phasing out of COVID measures. As demonstrated in the above examples, individuals remain hesitant to call for a roll-back of COVID control measures because of their innate fear of being accused of being careless/irresponsible on matters concerning public safety and health. Professor Yong also accepted that many of the adopted measures do not exist in silos. Therefore, it would be inconvenient to pluck individual control technologies out of the box as it would impact on the wider measures in place.
Adding further to the discussion on the role of “culture”, Professor Findlay highlighted that “culture” itself is a slippery concept. He reflected that Singapore’s culture is a melting pot between the East and the West and if one were to examine Singapore’s earlier lockdown and its discriminatory impact, it is evident that the lockdown discriminates based on one’s age, as opposed to particular cultural factors. Professor Findlay then stressed the importance of the concept of global citizenship. He recollected that the handling of bird flu took on a more global approach, and warned that the absence of globalisation would make it difficult to talk about the possibility of sharing resources such as potential vaccines. It is therefore crucial to question where each country sit within their global community. Professor Findlay concluded that we may all fall victim to poor governance and from nations' incapacity to share with one another.
Question 5: Contact tracing application models across Asia
Panel members were then asked their thoughts on which contact tracing application model is to be preferred i.e., centralised or decentralised. Professor Yong noted that the Japanese COCOA app that utilises Bluetooth technology has an adoption rate of 6% whereas Singapore’s contact tracing application has a 16% adoption rate. He pointed out that both countries’ take-up rate seemed significantly below what is required to make it effective from an epidemiological standpoint despite some privacy preserving aspects in its design. Professor Yong then queried whether partnerships with private companies should be struck so as to achieve a higher take-up rate? Or, whether authorise should consider adopting decentralised models.
Professor Findlay pointed out that the fascination with contact tracing applications stem from the assumption that it would definitively curb the spread of the virus. However, to date, there is no clear evidence to support this belief. He pointed out that even in Germany, where the application take-up rate stands at 85% of its population, the app is known to constantly breakdown. Hence, countries should be cautious that contact tracing applications are not used as the only pandemic containment method/strategy to the neglect of other useful tools. Professor Findlay then remarked on the position in Singapore. He pointed out that Singaporeans are typically unconcerned about their personal data unless it involves data surrounding their health, therefore, it has been difficult to secure the requisite take-up rate for the app to be effective. He attributed Singapore’s success to other methods including prioritizing manual tracing efforts and a QR entry code system. Professor Findlay further cautioned against the overreliance on a single piece of technology even if we can get the private and public sector working together. In fact, he warned that such partnerships may develop into new problems such as mass data sharing. He concluded that we must accept that it is possible that technology is unable to always keep us safe and it may be more useful to return to old-fashioned, manual support like testing regimes – a core strategy used in Singapore.
Professor Ema agreed with Professor Findlay that we should not simply rely on the use of one specific technology to prevent the spread of the virus but a combination of a few different tools. She then highlighted Japan’s Ministry of Health’s collaboration with Line, a messaging application. Professor Ema pointed out that when the partnership was first announced, the MOH and Line signed an agreement stipulating that the data shared is to only be used for the purpose of preventing spread of the virus, and is not to be integrated into other ordinary services. She highlighted that COVID-19 promoted the use of such schemes/partnerships because the relevant stakeholders can recognise that it is costly to build new technologies and more effective to work on existing data. Professor Ema then acknowledged that different barriers and structural differences exist in different countries and ask the other panel members what they believe makes it easier for their countries to build these structures and promote data sharing between B2B, B2C etc.,
Professor Yi Zheng responded to Professor’s Ema’s question by using the Chinese health code system as an example. He recognised that he had seen reports criticizing the system as being “unethical” but pointed out that “ethical” systems may not prove as effective/useful. He then moved on to say that regardless, one can still recognise the benefits of a centralised system – specifically, he communicated that individuals will know exactly who to hold responsible for when a data breach occurs – it is the government and not private actors who will be held to account. Professor Yi Zheng then explained the willingness of China’s citizenry to share their personal data with the government, he attributed it to the Chinese government’s proven ability to control and handle their personal data securely. He then highlighted how important it is for authorities to concentrate their efforts into securing these collected data. He communicated that it will allow people to feel safer and lead to less criticisms of the state powers. He acknowledged that it is difficult to strike a balance between building “ethical” and “non-ethical systems” but proposed that where a balance cannot be struck, the next best thing to do would be to keep the collected data safe so that citizens’ trust will not be diminished.
Professor Yong explained that Korea mainly utilizes QR codes for safe entry. However, he explained that many restaurants in Korea also adopt manual tracing methods by writing down customers’ phone numbers. Data such as home addresses and names are no longer collected for the purpose of contact tracing. Customers have the option of utilizing a safe screening app if it is preferred. Professor Yong commented that Korea appears to have struck a good balance by giving individuals the option to choose.
Question 6: Public reactions to state surveillance
The discussion then narrowed in on the topic of public reaction(s) to increasing state surveillance. Professor Yong Lim questioned whether COVID-19 is an outlier in this regard or will the public will become more accepting of these intrusions in the name of health security and safety? Professor Yong also asked panel members for their thoughts on ways to ensure that rights-invasions are minimised and the necessary checks and balances upheld/safeguarded.
Professor Yi Zheng communicated that the Chinese government has been alert to citizens’ reaction concerning its adopted technologies since the start of the pandemic. To achieve maximum citizens’ trust, the government would pilot test some of its measures and adjust its strategies according to public’s feedback. Professor Yi Zheng believes that a good balance is struck here by giving citizens “a voice”. He further emphasised that even if the imposed measures may sometimes collect sensitive personal data and breach on certain fundamental rights, citizens can be assured that the authorities will take the necessary steps to ensure that data is kept securely. Professor Yong Lim agreed with Professor Yi Zheng that this example shows the important role of civil society in balancing out policy objectives.
Professor Findlay expressed some scepticism with the above points and remarked that it is unlikely that centralised databases are retained by the state for a single purpose. He also communicated that it is problematic when states ask its citizenry to trust in its power and authority when there are no proper accountability regimes in place – while it is important to maximise protection and security for any data collected, it remains crucial to respect the right to privacy. Professor Findlay raised that it is untrue that we need to compromise personal data in order to be protected in the best possible way. He highlighted that certain countries like Germany and Norway did not violate personal data in its contact tracing apps and in spite of which, there is no evidence that points to the apps’ reduced efficacy. He emphasised that we should work off that basis and have certain expectations that surveillance tech must work from ethical and compliance design. It should not simply be left to the citizens to complain before the state does something about Surveillance apps should be built upon the right foundations that takes into full account constitutional rights and private sector agreements.
Professor Yong Lim then proceeded to explain the situation in South Korea. He communicated that government surveillance in the country is multi-entity and includes both the central and municipality governments. It is insufficient to simply put pressure on the central government because data leaks could also occur on the municipal and local governments level if mandated procedures are not adhered to.
Final Remarks:
Professor Findlay emphasised the importance of citizens’ inclusion and the connection between accumulated citizens’ trust and the efficacy of COVID control measures. Professor Ema encouraged further comments and questions from audience members. Professor Yi Zheng highlighted that everyone has to play their own role in society and the importance of citizens-government interaction in building trust. Professor Yong then proceeded to end the first session by encouraging vigilance and alertness.
Session 2: Life under COVID-19 - Changing Lifestyles and Governance
Moderator: Arisa Ema (The University of Tokyo)
Panellists:
Mark Findlay (Singapore Management University)
Yi Zeng (Chinese Academy of Science)
Sangchul Park (Seoul National University)
Question 1: Reactions to the first session privacy implications of the design of some contact tracing apps
The session started with Arisa Ema requesting that the panelists respond to the conversation held in the previous session and the questions that were coming in from the audience. Sangchul Park kicked off the discussion by responding to a question from the audience asking what the new roll out of contact tracing applications using Bluetooth meant for technology and privacy considerations. He pointed out global divides over centralized and decentralized models; as well as the choice between using Bluetooth or other forms of data like GPS or cell IDs. On the latter, he suggested that it would have been inevitable that people would be able to be identified. Concerns over privacy implications of such identification thus prompted many European countries, along with Singapore, to choose to develop their tracing apps using Bluetooth rather than other forms of geolocation data.
Mark Findlay built on these ongoing tensions by teasing out the privacy challenges of walking the line between anonymized data and personal (identified) data, the latter of which the state requires in the event of a positive COVID result. While lines might be drawn between both categories of data, these lines become blurred when different technologies mix: such as the use of both tracing apps and QR code for building entries—as is being trailed in Singapore. Doing so enables simple re-identification not only of people, but their associations as well. The pandemic has brought out instances of these elsewhere, and the consequences for marginalized communities can be dire (e.g., LGBT clubs in Korea or rising xenophobic sentiments).
Bringing in a Chinese perspective, Yi Zeng shared two stories about China’s ongoing experience with its health code. Firstly, when it became mandatory there were unanticipated social effects: old criminals voluntarily turned themselves in to the police because they needed the code to enter buildings and get around the country. Where they had previously chosen to live without cell phones, the institutionalization of the health code meant that this was no longer an option. The second story he shared was one that touched on the digital divide: while out with his mother, she realized she had forgotten to bring her phone along with her and was unable to show her code. The incident nicely captured ongoing questions about ensuring that the mandating of technology and applications do not end up leaving segments of the population behind. Thus, as Yi Zeng commented, as the government choose its response strategies, it needs to continuously ensure that corresponding infrastructures are simultaneously improved so as to ensure that no one is unable to participate fully in society.
Question 2: mobile infrastructure and trust
Does mobile infrastructure help with application of COVID-19 measures and specifically with building trust?
Yi Zeng then addressed question from the audience about whether the prevalence of mobile payments set the ground for trust in COVID-19 responses. He noted that Tencent and Alibaba may have played a significant role in establishing trust in the systems given that the two companies have shown their ability to control the data they collect and ensure its security, while also ensuring that their users are provided with services like their mobile payment capabilities. Given that the everyone already was accustomed to and already using these platforms, the growth of the health code within these applications made it easier to accept. On the other hand, he also noted that we might flip these expectations to ask about our expectations of data not collected by these companies. Will the government be able to secure such data? There have been instances in the past that services procured by the government led to the loss 2.5 million transaction details. Instances like this indicate that if existing infrastructure is absent or weak, then there are reasons to question the effectiveness and security of digital systems.
Picking up on this point about the relationship between the public, companies, and the government; Mark Findlay raised the question about the operation of big technology companies that tend to hold large amounts of data. He noted that the issue at hand was not about the size of the companies, but rather about transparency of data use. In Singapore, for example, legislation is applied differently between private actors and public agencies; while some have suggested that this may be necessary given the role of the government, it is worth questioning whether these rules should be the same across all forms of organisations. At present, data subjects are insufficiently informed about the uses of their data, and yet COVID has exacerbated the intersection of private sector and state bodies in ways that have never previously occurred. We may want information silos to be kept separate, but the pandemic has made data sharing a necessity: how do we best address questions of transparency and the ability of data subjects to have control over the uses of their data?
Speaking on South Korea’s experience Sangchul Park commented that the country was also confronting larger questions about the how to regulate big data and its sharing. This is a conundrum that the West is also facing, where the collection of vast amounts of data within credit bureaus have left these organisations vulnerable to security incidents. In South Korea, sectors such as the banking system is taking a centralized model in which small data holders lack resources, so relying on big data holder is inevitable. As a consequence, it is important to design regulation for banks and these big data controllers to manage the risk associated with these activities.
On the other hand, small data holders are also vulnerable to security threats because they underinvest in security, which seems to suggest that big data holders are inevitable but this poses its own set of challenges. We need to strike a balance but it is not clear what that balance can be.
Arisa Ema expended on this point of balance between companies, the public, and the government by sharing how the Japanese industry was often built on chains of business-to-business relationships, rather than business-to-consumer relationships. Long supply chains like these muddle the water of data sharing and complicate data governance questions.
Question 3: leadership opportunities and challenges for central and local governments
Arisa Ema turned to another question from the audience about the different leadership opportunities and challenges for larger countries where directives from the central governments may sometimes be misaligned with those of local and municipal leaders.
Commenting on Singapore’s perspective, which he admitted was too small to have such stark differences, Mark Findlay pointed out that the issue wasn’t about having a provincial responsibility network such as the structure in China or the US, but rather the issue was the extent to which neighborhood level organisations might participate in certain areas of control. In Singapore, for example, the government has introduced testing facilities within the country’s public housing estates which are operated for and by each estate. This has also been the experience of testing facilities for migrant laborers, where testing facilities have been decentralized. This experience of decentralization has been premised on the understanding that individuals are taking on more responsibilities, and that the country has adopted a multifaceted approach to allocating the responsibility for action. Nonetheless, issues of breakdowns between the federal and state governments in other countries continues to be a big source of tension, with significant ramifications for the delivery of medical supplies across these countries.
Yi Zeng used the opportunity to also respond to a second question from the audience asking about the freedoms that individual provinces may have to adapt their responses. He brought up the example of Hangzhou, which was the first city to pilot the health code. City officials had thought of also migrating this system to a personal healthcare system that tracks behavioral data such as sleep, smoking, and drinking habits—and then share this information with companies and institutions. Nonetheless, there turned out to be a large backlash against the plan which suggested that the health code and QR code system would be used even after the pandemic. The response prompted the local government to backtrack from their initial plans to build a more expansive system. The incident affirmed that individuals did not mind the necessity of the health code system within the context of the pandemic, but that there still needed to be boundaries that are not crossed. More broadly, Yi Zeng commented that many local governments have the freedom to try new designs, but are looking out to hear the public’s reactions before fully instituting new designs and changes.
Sangchul Park then commented on South Korea, which has so far benefitted from central and municipal government cooperation. Nonetheless, two challenges arose. The first was that as control policies were rolled out, it became clear that municipal governments had lax security practices leading to multiple identity leaks from within the government. Despite these cases being indicted, concerns regarding data security remain. The second challenge was that it was also the case that municipal governments would use the situation to promote or achieve populist agendas. He pointed out that these problems need to be addressed in order to ensure smoother cooperation between local and central governments.
Arisa Ema rounded off this discussion by sharing her perspective from Japan, where the declaration of emergency was first launched by Hokkaido rather than the central government. Over the course of the pandemic, it was often that prefectures and municipal leaders would take the first step depending on conditions on the ground.
Question 4: Individual and collective engagement in AI Governance, and the role of academic and professional networks
As the discussion drew to a close, Arisa Ema invited each panelist to share their thoughts about the panel and how to move forward with the issues of AI ethics, data protection, and AI governance that COVID-19 brought to the fore.
Mark Findlay touched on what the Asian Dialogues were meant to achieve, describing the series as an ongoing exercise to draw Asian scholars together so as to offer more perspectives in the ongoing global discussion about governance. He then observed that the pandemic has led to a massive upsurge in research interests like persona data protection, surveillance, civil liberties and rights. He also observed that all the countries represented in the discussion today have expressed a desire for developing themselves into smart cities, which are themselves developed via mass data sharing and driven often by surveillance technologies that are hard to understand and often insufficiently discussed. One consequence of the pandemic may be that more research could be done on the smart city: what kind of smart city do we want, and to what extent are we happy with having the private sector looking into our lives as part of the development of cities of the future? Connected to this is a question about governance itself: governance discussions have so far not kept up to pace with the pace of technology development, and so we need to be more vigorous about these questions.
Yi Zeng stressed the need to also talk about catastrophic risk more generally. COVID surprised all of us, but in the future, there are also other questions and developments on the horizon such as AGI and the role of AI. He pointed out differences between how the West have approached AI – which is often conceived of as a tool. Yet the discourse in Japan, China, and South Korea may be different. These countries often see AI as a member of society rather than just merely a tool. We need more strategic research which takes into account the long-term considerations and analyze different trajectories to secure us from long-term catastrophic risks like COVID and AGI.
Sangchul Park observed that there have been new opportunities as a result of the crisis—such as the penetration of smart devices across different demographic groups. Despite social distance, we are currently more connected than ever before, and as such we should leverage this change to transition into a digital area with proper governance. We might consider initiating talks on e-commerce, telemedicine, robotics, and fintech—along with talks that address social changes such as workers’ rights and digital divides. Discussions such as today’s enable countries to share their experiences so as to minimize trial and errors and speed up preparations for the coming digital stage.
Closing
Kideaki Shiroyama, The University of Tokyo
Professor Shiroyama closed the session noting the need to discuss governance issues are various levels: between short-term needs and intermediate issues (as when addressing the challenges of data sharing), which in turn influences tradeoffs that have to be made at different levels of government. AI governance also remains a key issue: what kind of society do we imagine for the future and how should we design a smart city using AI?
