Again, because of g0v participation officers, they brought this to our participation officer network. We did a full, radically transparent transcript of the three meetings to re-include that into the healthcare system.

On the experimentation part, I think people have to pay out of pocket. After people see the efficacy of this, then, of course, we can...Just like nutritionists, that’s another topic. A nutritionist during, I don’t know, post-ICU care and so on was not traditionally recognized by the Medicare system.

We are now in the beginning of the one-year sandbox application. Ming-yi is preparing the public experimentation plan as we speak. I think, started next month or next week, maybe, you will see the first cases that are legally recognized as part of the sandbox.

Many people who are suffering from mild depression or so on are socially excluded already. They are not motivated to walk out. It’s a lot of help if, when they are in that stage, they can still keep the connection with their counselors online, and the counselor can still be ...

We are in the one-year experimentation now. There is a thing called Ming-Yi that’s the company doing e-consultation. They want to prove to the society that it is actually as -- with the proper care taken, of course, on cyber security and so on -- that it is as effective, ...

I think it’s one of the best film, TV series that I have seen when it comes to talking about media literacy and things like that. It’s really good. In any case, yes, online counseling is now through the sandbox method.

Counseling is recognized as a certified profession, but it’s a recent thing. It wasn’t there 10 years ago. In this decade, people started recognizing the importance of professional counseling. I think the public TV film, "The Worlds Between Us," 我們與惡的距離, is really helpful in making that point across the whole ...

For these cases specifically, we use machine learning. There are many cases.

That’s what judges should do, but most of their days are actually consumed by lots of chores, which they may get ridiculed if they miss one of the evidences. There is no personal attainment at all if they do the sentencing right, because anybody would have done it right.

That’s right. It doesn’t require any value judgment from the judges. Of course, it’s limited only to the...Because then the judges find personal enjoyment when they can apply their personal value judgment to a case.

Now, this system does a machine learning and just automatically tell you what the sentence should be, based on the factors from the fact-finding alone. It saves a lot of time, and a lot of chores, from the judiciary and so on, and so forth.

It’s very, very mechanical, so no prosecutor feel a lot of...It’s just like the water hearing, skill masters. They don’t feel a lot of fun listening to the pipes that are not leaking. Neither do the prosecutor feel a lot of personal enjoyment when they just do those very formulaic ...

Then, for very mechanical sentencing -- for example, drunk driving -- drunk driving is essentially a formula. Whatever evidence you have, what measurement of alcohol in the blood or whatever, it automatically calculates the sentence level.

People don’t easily understand that. They crowdsourced a dictionary to translate them into laypeople’s terms. That’s especially important, because we’re starting to introduce a jury system for the first time. It’s for the jury to understand.

For example, in the parliament, people live stream their negotiation between parties. In the Traditional Yuan, they publish all the prosecutions, but the prosecution text, aside from ones that are confidential...In any case, the prosecution text is legalese.

It is a way to connect people. This one is very interesting. They used machine learning again to explain the public prosecution, because every branch of the government now adopt radical transparency.

80 percent of Taiwanese elders prefer to hospice at home, but 80 percent actually do this in the hospital. It’s a very large disconnect, mostly motivated by lack of communication between the primary caretakers and a special professioner.

It turns out they can detect it very early on using machine learning. It’s another very interesting one. This one is essentially making sure that people who care for the elderly can, in the hospice level, get real-time help from the different psychiatrists, from different medical doctors, and so on, ...

This one used machine learning to build a model to detect illicit financial flow for people who, for example, set up three shell companies to do business with each other, but actually are doing and using that credit history to get a large loan from the bank and then abandoning ...

This one is using climate change data to analyze when it is better to plant cabbages and things like that, to ensure a fair flow of information to the farmers.

That’s right. I’ll just introduce them in turn, I guess. This one is about a unified way, across different cities and counties, for people to report illegal parking and so on, using their phone.

Right, exactly.

Yes. Yes, the quadratic voting gets a lot of people’s attention, because you don’t have to propose a new project. You can just go online and distribute your votes, right? Also, we have what we call a wishing well, a wishing pool, so that people can just write their wishes ...

They are very eager to invite the team after the Presidential Hackathon to them as well. These are 20 very good stories, but I just mentioned two that work.

Now, New Zealand’s also very interested, because they have dairy farming, and who both require clean water from upstream and often get complaint from people downstream. [laughs] They really wanted WaterBox to help them measure E. coli or other common pollutants.

Everybody has the motivation to install the WaterBoxes. The Presidential Hackathon is, essentially, providing the blessing from the public sector to identify the most hotspots of such contentions, have people just deploy those WaterBoxes by themselves, and have a way to view a dashboard.

If they don’t actually pollute the water, then they are very eager to prove that they are not polluting the water. Again, the farmers really want to prove that which upstreams are responsible for the chemicals in their farmland.

Now WaterBox is particularly useful in Taiwan, because what Taiwan has is a lot of industries really close to farmlands. Now, if they pollute the farm waterway, we just passed a law that can just unplug their electricity and water.

They automatically report it to the Internet. This team also did the AirBox, which is very successful here, to measure air pollution. I talk about it at Republica, and they use distributed ledger to make sure nobody can change each other’s numbers, as a way for social solidarity and trust.

This year, there’s a very similar, also water-related, where people use very cheap, inexpensive, boxes to build what they call WaterBox. Every part of it is made in Taiwan. What it does is that they can be placed into the rivers, into waterways, into all the different places, and measure ...

It really is very significant diplomatically, because unless we trust another country, you wouldn’t hand your water pressure and water flow data to that country’s Presidential Hackathon team. The New Zealand people really trusted us, and it is a very fruitful collaboration.

It’s so good that New Zealand, Wellington, discovered this team right after last year’s Presidential Hackathon and invited them to Wellington for another three months.

They just go to those places, and it has more than 70 percent chance of getting it right. It’s now reduced to a couple of days to discover the leaking point. It’s very useful for machine learning, and it doesn’t have any privacy impact. It’s a really good model.

They build a chat bot online, actually, so that masters, the skilled craft person, can wake up and see a robot apprentice who will report every day to them what are the three or four most likely leaking points near them.

Right. Basically, the machine learning tech experts, as well as policy experts, look at the water flow and pressure data that’s already installed all along the pipeline, and then they just did a machine learning model to detect unlikely new spikes in water use.

They employ people who are very skilled to listen to the pipes to discover whether they are leaking. It takes, on average, two months to discover a new leaking point, because most of their time is just squandered by listening to things that are not leaking. It’s not a very ...

Right. Last year, a winning team is called Water Savior, because they save water, you see. The solution that they developed is basically to solve water leakage. In Taiwan, the Taiwan Water Corporation maintains one of the longest plastic pipes anywhere in the world.

The presidential award will be next week, so we will have two winners from the international team, and five from the domestic.

This year, we also have an international track. We got requests from, I don’t know, 16 countries, or something, and now they are in the final seven teams. Next week, actually, I’ll be meeting them tomorrow.

Yes.

We delivered on all five of the last year’s Presidential Hackathon. We now have a lot of street credibility, and people are very eager to participate.

We may change 12 regulations for it. We may allocate millions of euros for it. We may do personnel change because of it, but whatever you do, we will do whatever we can to make sure that the top five teams become public policy, starting next year.

After that, if your director-general doesn’t like your idea, you your project your president, and he, or she, must say yes. It signals the presidential promise that the five teams, after three months of prototype, within the next year, we will adopt them into public policy.

If you’re a public servant, you participate into a team, it’s just a hackathon, right? It’s three months, just weekends, so there’s nothing to lose, there’s no risk. If it only gets people’s attention and it finally gets people’s contribution -- I can tell about some specific cases -- then ...

After that, the president gives a winning trophy to the top five data collaboratives born out of these 20 teams. The five teams basically get a trophy, that there’s no prize money. The trophy is a projector that, if you turn it on, it shows the president handing the trophy ...

This results in a very balanced -- because people look at the synergies between projects -- a very balanced cohort of 20, and each of them, we coach them with mentors and so on, to make sure each of them become trilingual, that, say, each team has a policy expert, ...

This actually ensures that nobody will just dump all of your points on a single issue, because it’s not technically possible. 99 is just 9 votes for 81 points, and you still have 18 left. You have to think of another one to have four votes, at most, and so ...

Basically, starting with 100 or so teams, everybody gets 99 points. With the 99 points you can vote any of the teams, but one point gets you one vote on a team. If you want to add to that team, for two votes, then two votes is actually four points, ...

This year, alone, it’s 110, or a similar amount of teams. It’s a huge number, and we also use a new voting system called quadratic voting and asking everyone to select the cases they want to see online.

What I am seeing in Taiwan is that first, g0v grant took the prototype grant model, and then we also look at g0v grant, and lifted it into the presidential level. Now, Dr. Tsai Ing-wen, our president, every year, looks at many, many different teams.