Audrey, you were mentioning earlier about the new Ministry of Digital Affairs. Tell us about that. What’s the plan?

The “moda”, or Ministry of Digital Affairs starting this August or September is a new ministry in the Taiwanese national government in charge of digital affairs. We structured that as an institutional organization of the past five and a half years of me working in a cabinet office as an at-large minister on digital affairs.

It turns out there’s four pieces of state organization that interacted the most, including the Department of Cyber Security, including National Communication Commission on resilience like low-earth satellites, networks, and connections, resource allocation and so on.

And also, the data-related platforms to open data, MyData, and shared data, stuff from the government digital services that’s originally from the National Development Council.

Finally, the parts of the Minister of Economic Affairs, our platform economy, our software industry, things like that. We’re taking all these different pieces together and call it a new ministry.

That’s hugely exciting because that’s surely a real recognition of everything that you and the team in the community really has achieved over the last five and a half years.

Yeah, there’s that. There’s also the recognition that we’ve moved towards the second half of our original plan, the DIGI⁺ plan.

In 2016, when I first joined the government, we have this plan for a digital government called D-I-G-I, which stands for Digitization, Innovation, Governance, and Inclusion.

It starts from the more the e-government’s level of digitizing our services, to promoting social innovators, public entrepreneurs, and so on. Then, we’re now tackling the far more interesting issues of cross-border data governance and service governance, and also how to empower our 20 national languages as part of the inclusion pillar.

We’re realizing that the second half of the governance and inclusion pillar, it’s a lot like the Ministry of Foreign Affairs, in the sense that MoFA, the Foreign Minister, doesn’t regulate foreign countries. Certainly not as a way to hand out licenses for other entities overseas to exist, but rather finding common grounds, common values, and based on shared values, deliver innovation that works better for everyone.

We style the logo of “moda” as entirely a lowercase and intentionally refrained from taking a licensing role or a regulation role. For example, this spectrum day-to-day regulation management is still in NCC. The handing out of banking and purely digital banking licenses, they’re still in the Financial Supervisory Council. We’re a regulator and licensor of none, but a promoter of many.

Nice. Presumably, that’s only possible because you’ve been able to really build a thriving and highly capable a really mature digital ecosystem. Do you think what you’re describing would have been possible five and a half years ago?

Not at all because if we did not make the digitization equitable, if we did not deliver broadband as a human right, our educational competence, not just literacy education and not just basic education, but also lifelong education… If we have not demonstrated that it is actually possible for the civic tech people to set the norms around contact tracing, availability of PPEs during the pandemic so that we fight it without lockdowns, then we will not be able to have the legitimacy or to political capital to say now we’re looking at the other issues.

We are reasonably confident that we can also do a what I called a people-public-private partnership, where the social sector assess the norm, the state amplify the norm, and the private sector cannot help but implement according to the community and the social needs in a position of this over-centralized decision power stuff.

Fascinating. For some digital units, it’s a real moment of risk though when they move from being either within a prime minister or president’s office or as a transversal program with a roving brief in their own ministry. Is that something that you’re concerned about? If so, what are you doing to mitigate against almost becoming a silo in your own right?

There’s two parts of answer in that. First, we’re still playing an empowering role. For the deputy ministers, that’s the CIOs of the other ministries, we are not suddenly giving out orders or vetoing their plans. In Taiwan, we’re running with a EU-compatible, GDPR compatible privacy act, for example. The only thing that blocks us from getting full adequacy is that we don’t have a single DPA.

In essence, the health minister and his CIO is the DPA for things that are related to health and the transportation minister for things related to transportation. There are like a dozen different norms around data use, around privacy, and things like that.

The “moda”, the new ministry, definitely does not take a national DPA role. What we are doing is still empowering the different ministries in terms of capacity building, maybe lowering their transaction costs.

Instead of solving this N-times-M issue of data connectivity, we have this T-Road, which is like the Estonian X-Road, that makes it easier for them to write and adapt it once, and then have this full, reciprocal transparency everywhere and so on.

We lower the risk and lower the cost for the other ministers, but we are definitely not saying that “now this is my budget.”

That’s very interesting. It’s almost like the idea of servant leadership exemplified as a ministry. That’s amazing.

Exactly.

You mentioned capacity building. That’s something that’s come up a lot during this event so far. What kind of lessons have you learned in terms of what works in building digital capacity, both within government but also within the broader ecosystem?

Because Taiwan did really well during the pandemic, yes, with an essentially net-zero excess mortality, so a lot of boom in our economic sectors. That created a real problem for the state to retain digital talent, because the TSMC, the semiconductor company, along with everybody else, is essentially offering three times salary on pretty much all the positions.

That did make capacity building challenging in the past couple of years. In designing “moda”, we took two innovations. First, we said that the time you spend on, say, open-source projects, counts toward a diploma.

Say I’m a junior high school dropout, dropped out when I was 14, but I spent decades on open-source community. That earns me like five PhDs. [laughs] The point here is that the salary which in the Taiwanese government used to be very much diploma-based, is for the first time now based on your actual community contribution to open-source and other technology communities.

The TSMC still have not competed to that, so it gave us relative advantage when it comes to people who do not have a PhD but are well respected as senior leaders in the open-source and open-government communities.

That’s one. The second thing is we’re also offering a kind of prize, award for resilience in cyber security architectural contributions. If you make an architectural contribution that saves X amount of time, and energy, and loss of cyber-security-related assets and so on down the line, then you actually get a fraction of that as a prize, as a payment, and so on.

It encourages people to think about cyber security not as something that’s purely reacting to things, but actually designing the zero-trust parameters using the latest zero knowledge, homomorphic encryption, and so on.

Giving a real incentive for the technical people to be involved in the early conversations about business logic as early as possible to promote privacy-enhancing technology, the algorithm, and so on, by associating it with a real contribution on the cyber security era, which is also a very hot topic now with the geopolitical turn of things this year.

Of course. That prize is available…

Nationally for anyone working in cyber security and digital-related positions. If they make contributions either mitigating an attack or architecturing shifting into something that’s more resilient to attacks, then there is a prize money.

You mentioned some of the devoted geopolitical trends. We’ve just had a conversation here at the Bellagio Center around some of the big trends that will shape our world over the next 10 years and how they’ll impact digital government.

What is another one or two trends that you would pick out that you think will be particularly important for Taiwan?

I think the zero-sum politics is the trend that I was the most worried about, yet it is kind of the value proposition of Taiwan, because we’ve very successfully made open government, social innovation, youth engagement, these topics, literally all the four parties in our parliament signed on the open parliament deal.

They compete on being more democratic and more open, because that’s our founding value. The point here is that back in 2014, before we occupied the parliament, there’s this general apathy. People just feel that politics is just tribalist and it’s just zero-sum. It’s not worth the young people’s time to invest into thinking how to do digital services better, because the private sector, which does not have this tribalism, zero-sum stuff, by default deliver things better.

I see a real connection between this political apathy, this assumption that it’s zero-sum politics, vis-à-vis that the digital pro-social media, the spaces that we create, the digital public spaces or digital civic spaces that can then foster this kind of co-creation, making a non-partisan issue.

When we resolve, for example, the Uber gig economy issue or the Facebook paid advertisement before election issue, many, many other issues, disinformation issue in general, and so on, without walking a partisan line, literally with all the different parties supporting the civic tech, the open government to work.

Then it attracts the best talents, maybe already spend some time in our vibrant startup systems with some web3 exposure, a lot of Ethereum researchers and developers in Taiwan, who then turn their research idea they learn from the web3 startup world and say, we can actually use quadratic voting in the Presidential Hackathon, the president actually rewards this kind of new governance systems.

Democracy as a form of social technology that goes beyond the zero-sum politics is really deploying, and I’m really happy to see that many people also identifying, going beyond left or right, or any sort of polarization as the deliverable for the framing of narratives of a digital government.

Fascinating. Is there anything today that folks have said that has surprised you in terms of trends or other government digital teams are thinking about?

Yeah. I think this term of the Estonian technology stack, which we always look up to, suffering a midlife crisis, that really made an impression on me.

It really is true. What was on the Web 1.0 days, the best design for modularity, for consent, for reciprocal transparency, open-source, and all that, it doesn’t necessarily mean that the young people, which is growing up on a very different technological stack, will naturally find alluring or wish to continue to maintain forever.

A lot of the young people, they do not see themselves as coders, but because of the local no-code movement, they were able to piece together very compelling digital services.

They don’t think about code anymore, but this very coding-oriented, informed consent, inspections, co-creation by the open-source rules, pull request, and so on, is essentially a form of democratic governance over the algorithm that runs the society, but is increasingly seen as a legacy by the younger people who just snap their finger and put the services together and so on.

There’s difference of generational expectations by still very well meaning people. This is surprising to me. I haven’t heard the Estonian experience being referred to as a midlife crisis.

It’s something it seems like many digital service units that particularly are 10 years old are experiencing in different forms.

Yeah.

How do you essentially keep renewing the capability and motivation to keep that sense of purpose while also dealing with what is essentially increasing technological legacies…?

I think we got it easy, because at the very beginning, I’m the 數位政委, the digital minister in charge of Open Government, Youth Engagement, and Social Innovation. These are the areas, as I mentioned, that enjoy cross-party support, and built in in itself a sense of renewal, just because of the topics.

It also helps that digital in Taiwan, the word shùwèi, also means plural. It literally means “more than one way” to do it. So I’m also the minister of plurality, meaning that digital is never replacing paper, replacing labor, replacing public servants, but it’s always about even more inclusive, taking care of more national languages, making it more easy for the elderly, for people with different abilities and so on, to access the same services.

There’s a lot of inclusion-based values being built in, and it’s assumed that digital is here so that we can collaborate across vast differences. When the initial value proposition is like that, there’s less legacy, because we’re not attached to particular gov stacks. If the civic tech people invents a better way to do contact tracing, then overnight, we switch to that.

You’re more focused on the way of working, collaborating, and organizing the technology.

Very much so, yeah.

Just coming back to that notion of plurality, one of the things we were speaking about earlier was different views and more positive visions, if you like, for AI. Could you talk a little bit about that?

Sure. There is this singularity view on AI, where the AI essentially concentrates everything there is to know about one particular line of work, or about how people run their lives, and so on, and therefore forecloses future possibilities by essentially all the decisions and automating away those decision-making places in the society.

The point of the narrow corridor is to see that over-concentrating this authoritarian power, although it seems efficient on the short-term, is essentially giving up on future possibilities, because there’s a de facto lock-in. It’s actually the most fundamental lock-in, because there’s no competitor. There’s no room for competition anymore when you just cede control to automation.

On the other hand, we’re not saying that we should not collect data, or build data models, or things like that. What we’re saying is that it should be assistive, in the sense that my eyeglass is aligned to my best interest, my vision.

It’s accountable. If it expires, if it blurs, I get to fix it myself, take it down the street so anyone can fix it. I don’t have to pay five million dollars and sign an NDA. Basically, it’s aligned to the social norms, and it’s not about privacy, although, of course, if my eyeglass start popping out advertisement, I would say it’s very intrusive. [laughs]

It’s not doing that, because there’s an underlying social contract that says the people who make eyeglasses are just providing reference implementations. The people wearing eyeglasses end up doing the augmentations they need.

It’s a humbleness in the design that it’s not taking my agency away, and certainly not stuck or addicted to it. I can take it up any time. It’s not sticky. Also, I get the final say in terms of augmentation. Of course, this was not just me, but people who wear eyeglasses in general, like a user community.

If we build AIs based on the line of thinking of assistive technologist intelligences, then human dignity alignment already goes without saying. The real trouble begins when we start treating humans as fungible workers that’s measured purely on the per hour output, and things like that, which would then build a case for authoritarian intelligence in automation. The end result is indistinguishable from the authoritarian surveillance state.

Right, exactly what we were talking about as being one of the least desirable scenarios in this ideological model.

AI as automation is an ideology that’s actually isomorphic to the surveillance authoritarian state.

How do you make the case for your alternative vision of Assistive AI? Who has to be convinced?

I think, mostly, it’s about people who are in the policymaking or giving out grants to ask for some very basic things, whether there’s technology and power and users to communities. When people want to say, “Let’s look like these data models so that our school can augment the model and fund it very differently.”

The AI company or the AI project insists then their audit data gets aggregated to the central government, to the state, or some multinational organization, or know we are adopting a federated way of learning, which is sharing algorithmic insights, not data.

It gets back into the community of practice their choice of models. It’s two levels. One is at a decision-making level of giving out grants and research money. Also, in the education level, not locking in the students to particular ways of thinking about AI, but rather as a model of free choices that they can actually piece together their own models by remixing, just like remixing each other’s scratch projects and games and in a very low cost or cost-effective way.

I could go on all evening, honestly. [laughs] Maybe just to finish with, is there anything else that’s top of mind for you right now either sparked by state or just in your role that priority that you want to share or talk about or highlight?

Yeah, I really liked the Narrow Corridor framework. My main worry going into this conversation was that us in the democratic world may actually diverge in a sense of like, for example, in the US, for certain basic services, it’s deferred or ceded to the private sector too much, that acceptable corridor does not overlap with the rest of the world.

Upon conversations, it seems like it’s not the case. There is a real wish for the state capacity in a way that is, of course, still checked and balanced, that provides some, I would say, openness. Openness for, not just augmentation or remix, but also openness to future service innovations and something as basic as filing tax. [laughs] The experience of filing tax, I mean.

The filing tax experience was actually the first experience that we co-created with civil society back in 2017. That’s turned something very painful into something that’s quite beautiful. Just use your phone. It’s less than one minute for most people, and it’s a cheering experience [laughs] for most Taiwanese people. This shows the value of co-creation.

First, it’s good that the corridor seems to be overlapping. Also, the conversations, especially around the Ukrainian experience, shows that the corridor could also be widening.

There are parts of private sector contributions, that in a emergency response framework, the private sector is also happy to contribute to the state in need, in this case, Ukraine, to use it to further their purposes of public service without asking any lock-in. During a war, you don’t ask for lock-ins. That’s how wide that’s their corridor was.

Also, experiences from India, from other places, also shows that what used to be considered leviathan-like moves, ID cards and things like that, once tested in court and when other jurisdictions also think deep and hard about it, of how to expand the social sector capability to oversee this kind of expanded state capability. It also opens the corridor leftward, in that what seems like an authoritarian move could ultimately be checked and balanced, and maybe moved back into the civic tech framework world.

These two very welcome developments widens the possibility of a corridor, meaning that we can continue to work with both the seemingly more authoritarian state moves and seemingly more capturing private sector moves, but with this shared protocol of not foreclosing future possibilities of co-creation.

There is a real plurality between the state-provided services and private-sector-provided services that can communicate with the community-run or social-sector-run services.

That is a lot of people actually mentioned the DLT, the web3 community, and so on, as one potential target for engagement. If we don’t engage them, they essentially self-regulate and go into another world.

And almost other instincts…

Certainly alternate sovereignty.

It’s a corridor or somewhere else…

Like hyper-dimensional.

If we learn from their governance experiments and turning that research into development, then we meet somewhere in the middle, because these are the people who are interested in doing govtech. It’s just their government is not Westphalian.

Engaging these people on their terms, not on the terms of state capability or private-sector capability is also one very interesting thought to expand the corridor in.

That’s really interesting. I think even just lately, some of those communities govtech is an interesting framing on how could you potentially harness that. Fascinating.

Have there been any other examples that you’ve heard of from other countries that have really inspired…What in the last year or so has really caught your attention?

I’m quite impressed by the Japanese example of setting up a digital agency, rolling out the My Number system of national identification numbers in a really quite Taoist, low-key, servant leadership framework.

Like the moda, the Digital Agency in Japan is not taking over METI, their economy ministry. The DA is not taking over any regulator role. It just wants common protocols to make the public service life easier. It’s very humble, but it has a very strong founding value of “leaving no one behind”.

Literally leaving no one behind, just like the Taiwanese position, we’re now focusing on inclusion as the end state, and governance to further inclusion, innovation to further governance, digitization to further innovation. That is the right contrast to solutionism.

Just 10 years ago, a lot of places were focusing on the digitization, innovation, are using the smart city narrative, like “smart city, dumb citizens, but maybe that’s OK?” [laughs] Now, it must be all about smart citizens, and citizen empowerment, things like that.

I think the sustainability framework, the idea of co-creation, governance, and so on, were marginal ideas 10 years ago, but in the case of Taiwanese MODA and the Japan Digital Agency in particular, become the most prominent value-based on ideas.

I wonder if we can just pick up for a moment on inclusion, given that it’s one of your focuses coming up.

Yeah.

We’ve heard a couple of really interesting things today on that. One is this idea of inclusivity is a constantly moving target. Of course, as technologies evolve, that gap could widen in all kinds of ways.

Also, we’ve heard this concept, particularly from India and Bangladesh of inclusion through an entrepreneurial approach to bridging the last-mile gap, if you like, and being able to access and use public services. I wonder, what is the frontier for you, if you like, in Taiwan?

In those examples, it provides a real social entrepreneurship incentive. The state will not be saying that we compete with the local private-sector provider, which is usually something that democratic state doesn’t do. We don’t turn things into state-owned enterprises.

Precisely because for those last mile, there was really no case for the business that operate at debt. So when a state go ahead and do that, it’s just a reference implementation. It makes future entrepreneurship possible. It’s social innovation leading industrial innovation in those sectors, and I think this is a very powerful thought.

In Taiwan, when we use the polis system to deliberate online and also offline about how UberX should enter Taiwan, that was actually one of the consensus point. In that, of course, people agree broadly that they should not undercut existing meters, should have insurance, and so on.

One of the things is that Uber, the company, is actually proving that the platform for dispatching cars effectively could work in the least accessible places if operated by the local communities themselves, by local temples and churches, and so on.

One of the consensus out of that we have in our conversation is that whatever Uber is enjoying out of our newly adjusted Multi-Purpose Taxi Act, the same Act must also enable exactly the same platform cooperatives out of the least served communities precisely because Uber will not operate there. [laughs] It’s not profitable.

The local, for example, caretakers of long-term healthcare, once they gain a professional driver’s license, benefit from this new Multi-Purpose Taxi Act that we co-created because they can then deliver transportation service, not just for the people they care, but also for other people in the same town and also at weekends as tourism guides and things like that.

That was not possible before the Multi-Purpose Taxi Act. Uber is now a legal taxi company in Taiwan, the Q-Taxi, but there’s also local co-ops. One that’s just bringing up as a result of that co-creation. This is actually a brilliant example of what I call the people public-private partnership is the passengers and drivers, including Uber drivers, setting the norms.

When the state amplify those norms in a democratic way, Uber cannot help but commit to those norms, otherwise, they face strong social sanction. We apply that model in many places by essentially saying what’s disruptive [laughs] in Silicon Valley must actually be community empowering the plurality Taiwanese are seeing.

Amazing. That inclusivity lens is coming at a very early stage, almost at the policy level.

Yeah, at the agenda-setting stage.

At the agenda-setting stage.

It must be top on the agenda.

Right. It’s not something that you could do presumably post-hoc.

Probably not because if you regulate taking only the existing taxi association and Uber’s interest in mind that their common interest is to drive [laughs] away this community coops.

Right, exactly. Fascinating. Thank you so much.

Thank you. [laughs]

That was fascinating and hopefully, we have plenty of good content, I’m sure, that we are able to share.

OK, cool. Thank you.