…The before part, which is helped by… There’s a slide that talks about the SMS, the trusted SMS number, right? The 111. This is a new invention this year that lets us send out random SMS messages to random people and to hundred thousand at this time. And it went to 170,000 numbers successfully. And so, when people are filling the survey attached in the SMS, first they know it comes from the government. They can trust it and also we have a much wider sampling than traditional like rolling wave or things like…
Because one of the big, like where this kind of segues into social capital theory for us, is that, you know, you see the bonds of trust crumbling all over the place, except in very small groups in this country. So family groups, community groups, those. And that, you know, that for me feels like a source of hope, but also a source… like, to not be kind of starry-eyed about that, because it’s also about atomized communities.
However, because across the screen, you cannot abduct me anyway, right? So, the harm is limited across the screen. And because of that, people come to trust the people they just met across the screen and talk about the most intimate and personal things because they know that it’s harder across the screen for them to hurt each other. Of course, that was before the social media and cyberbullying and broadband. I was talking about the early internet, right?
The end goal is that, whereas we see anything CMMC compatible or NIST-approved as fit for integrating to the Taiwanese zero- trust posture, we want the reverse as well. And we do have our own certification capability. And I think that will massively increase the investment from the private sector, especially the semiconductor tool chain, to cybersecurity. And it will also make the de facto standard established by US and US DOD something that’s more international as well.
As a digital minister, my workspace is totally, radically transparent. All the meetings that I chair, the full transcript can be found on the Internet; it’s entirely published. By radically trusting people and making the state – and how the state works – transparent to the people, everybody can understand not only the why of policies, but also the why of policymaking: the context of policymaking. This is as opposed to other jurisdictions which make the citizen transparent to the state.
I think that is also that many people felt as well. For example, Ethan Tu, when he returned to Taiwan from Microsoft, as director of Cortana and speech technology, he’s basically running a charity here. He’s running a foundation called AILabs that work with the local AI people, and try to find out how AI improve the society, in trustworthy way. Again, I am sure that he took a huge pay cut, compared with his work in Microsoft.
It took two years for me to be generally trusted by the senior career public service. And after I become the minister in 2016, we elevated that system to a cabinet level Youth Advisory Council. So now we have like 30 or so reverse mentors as a institution, chaired by the premier, and co-chaired, I think, by Minister Lin Wan-yi now, basically they work with all 14 or so ministries, each nominating one or two young people under 35.
I actually like that model. So separate repo for talk plus features. Because I am very interested in playing with those features, but they are not necessary. In fact, it’s likely a turnoff if it is very priority. Because if you say I don’t trust the AI to talk to it to get a conclusion, that shouldn’t be an issue. Anything else that comes to your mind that you would say this would be good to have?
It’s really cool, it’s like, with just a whole bunch of peer-to-peer connections, you’re able to calculate people’s global trust score. And this seems like really… like this is very researchy at the moment, but it seems like a very exciting idea to be able to, once we have this data set, it’s really nice to be able to… you can have all these different ways of thinking about the data inside the set.
We’ve been talking with Zoom about this. We’re glad, of course, that they introduced, for example, end to end encryption, although it’s not by default. But at the moment, for example, in a web based Zoom conversation, it’s still not supporting end to end encryption, meaning that we either have to use a device like iPad or a phone or something, or we have to trust its web interface on a non-end-to-end encrypted form.
I’m curious from another historical perspective, how the Taiwanese government hacker community of which you were part reflected or responded when we had the Edward Snowden revelations and the greater awareness of the American National Security Agency’s, tactical operations, which involved trying to weaken thetrustworthiness of hardware coming out of other parts of the world, including the US. Does the same norms that you talk about, trustworthiness with respect to PRC, apply to other states, operators, and other countries?
That is the thing we have there, and otherwise, it is just a mess. Switzerland is a mess. It has a reputation as a stable and great democracy, and people vote four times a year on a federal level and stuff like that, but there are no real standards, not even for the paper voting area – in a way that you can always have trust . But of course, if you want to rig the whole thing, you need a big conspiracy.
As part of ROOST (the Robust Open Online Safety Tool), we have been working with frontier labs to create open models. One released just last week was the safeguard model from OpenAI. It is the first reasoning model for trust and safety judgments that comes with a full reasoning trace. The idea is that this model, which is small enough to deploy on a laptop or a community server, can ingest communal policies. It is “Bring Your Own Policy” (BYOP) .
So socially, as long as there’s general societal trust , fact checking does work against information that is blatantly false or can be shown to miss important context, but to meddle in the election or to get people to buy something or whatever, to do persuasion. One does not need this kind of factually untrue messages. One can just trade on factually true messages, but just with the highlighted importance of things that increase the effect and the mobility, like going viral.
This innovation works really well with quadratic funding because previously, the panel of judges was either a few people that the government trusts a lot — but then maybe they’re not diverse enough — or it’s just crowdfunding — but then people with a lot more money actually speak louder almost by definition in the crowdfunding scenario. So quadratic funding is between a panel of award committees and pure crowdfunding, and chooses something in the middle that has hopefully the benefit of both.
There is no guarantee that a so-called private sector entity we’re dealing with will not become state-owned the next day. This is too much a risk. The second is path dependency. If we use it for a system, then it’s always easier to keep using the same supplier, 3G, 4G, 5G, and for all the upgrades, and all the natural disasters or whatever service outage, you have to trust the vendor to come up with emergency hot fixes.
Absolutely. There is a very delicate place where there is this maximum tension between the protester/demonstrator movement that is gaining attention and really pushing the political leadership to its limit, and the political leadership seeing that they lost power and the trust of the public. So what do they do? In most situations, violence happened to repress this new power that was coming up from the people, right? But in this case, they didn’t try to co-opt the movement either.
And for an information society, that’s shaping its core, really. So, there are some efforts to more protect authenticity and attribution or have new media formats, and that’s my domain. I see there are some systematic approaches to attribution and data provenance, but I don’t see it implemented in that scale. Maybe first for news and trusted information sources, but I don’t see the roadmap for this right now, at the pace probably we would need it right now.
What they lack is the empowerment that amplify their collaboration models to the national level because of communication barriers. It takes forever to communicate the specificity of a local. If you want to work on community building for example, it takes average five years until the local people trust you. [laughs] It’s hard to just port to adapt like this but what we’re doing essentially is making sure that people who have a good social innovation, we absorb all the risk.
It’s OK. This is what we did. We set out a projector like this. This is the occupied place, it’s the parliament, but there’s a stenographer, three actually, rotating. There’s a stenographer in the parliament sitting here, typing everything she heard. Here, in real-time, a stenographer. You don’t have to trust what she writes, of course, buy anybody with a phone, or a headphone can check whether she’s actually being factual. She has to be factual.
