…And because we have built more than 100 bridge-making experiences before the pandemic, people generally trust the government to deliver this kind of uncommon ground in a timely manner. So in 2020, the approval rate of the Tsai Ing-wen administration went through the roof, like more than 70%.
…As some of you know, many leading countries have joined together and sign the global declaration, information integrity online, which is defined as an information ecosystem that produces accurate, trustworthy and reliable information meaning that people can rely on the accuracy of the information to access while being exposed to a variety of ideas. And this is a direct result of this generation of Frontier AI that can convincingly synthesize people, whether it’s text or image or sound or video and so when it leads to a lot of deceptive AI content online.
…But things like cyber security, they’re often confidential, often national security matters, right? So I set up this policy that I don’t even look at them, so no confidential or top secret information pass through my eyes, which is why also I can not give you an answer, because most of the cyber security really matters are classified, and I don’t even see classified information, but I do trust Howard 簡 and taking care of these matters.
…One is the trusted institutions and numbers. One is the friends and families who have met, and everybody else is a bot.
…My main work at Apple or at the Oxford University Press is, actually, introducing them to works from the open source society, and people who reached out to me from those organizations — such as Socialtext — were already friends that I trusted . They are fellows in the open source and free software movements. Their position was exactly like how I worked with Taiwan’s Nationalist Party government in the past year. Everybody expects the Nationalist Party in Taiwan to lose in this election, but still, over the past year, we worked with the administration to…
…And as you observe this, what do you think that format of that governance looks like? Is it that we end up with some trusted corridors between regions and some untrusted patches of the internet, or is it that we end up with a lot of closed national internets and you have to do some really close checking as data comes into your autonomous area? What do you think it ends up looking like over the next couple of decades?
The point here is that instead of saying, “The young people have demands and the government should please them,” it flipped it on its head and said, “The young people have directions and the senior people should follow them and provide them with resources.” That turned out to be the Pygmalion effect; it was the only way to restore trust because the young people really did have better ideas of where to go. The career public service then started to see those young people who joined as reverse mentors as comrades who were in it…
…So for these very privacy oriented issues already, our private sectors are trusted partners, and they are, in a sense, taking over with some public oversight. And with time, I think more and more, as long as the social sector feels that they can exercise the same sort of control and auditing to the private sector, we’re seeing more and more of this role shifting into the private sector.
…So people would get video calls from people they trust and asking them to invest in a new stock or things like that, because the Taiwan stock Exchange is doing really good. So there’s this new breed of famous celebrities just start video calling people, which is one of the most detrimental to the society uses of generative AI.
…So from 9% to more than 70% in a short span of six years, just by trusting the people.
…He’s not shy from just delegating all the government’s data, and its power, and everything previously invested, to the public and to the civil society. The other unique thing about Taiwan, is a collective priority of rebuilding strong mutual trust between the government and the civil society. This is something that I think a lot of European nations, especially people who worry about the private sector having too much control of the government, could perhaps look into, and then have more conversations with Taiwan for this kind of collaboration.
…It should be trust -under-loss: whether people, even when receiving an unfavorable outcome, can still accept the process as fair. Social‑platform algorithms are a case in point. Publishers who want to broaden readership may work on better headlines and timely angles. But machine‑learning systems boost controversial posts, amplify ad hominem attacks, and surface baiting content—because the machine, absent ethics, chases traffic. This single‑metric optimization around “engagement rate” ultimately corrodes the quality of public discourse and mutual trust . Take the Netherlands again. Before cutting a family’s benefits…
…This is about this idea of a shared cognitive, shared reality, fabric of trust that these emerging technologies employed by authoritarians.
…It’s just a Post-it Note to remind oneself where have you been in the past four weeks, so it should not be wiretapped.” And so after we did this explanation, again, the public trust of the system increased especially because people do see that within 24 hours, we do send out automatic exposure notifications that people can then quarantine themselves, and so on. So radical transparency refers to the entire configuration, not just one or two players in it. It’s only if everyone is mutually accountable, can we actually do this large-scale…
…And so this pollution to fabricate trust is wide ranging. It is not one operation. It’s that each operation decimates. Just like Freon decimates the ozone. Right. It decimates the epistemic commons. So that’s my second point. So, to counter that, I think we really need to go back to the actor, to the source. All the governmental utility bills, electricity, water, whatever, are sent now from a single number, 111 in our SMS and all the three telecoms to their commercial SMS services have all adopted short codes.
…Like NIST or Cybersecurity Institute and so on, we can play a trusted hub like the CVE. Or anything that the frontier models can share the incoming vulnerabilities without compromising trade secrets and confidentiality. This is something I’ve talked at length with the director of NIST and other similar institutes. We all think that even though it requires more expertise than traditional cybersecurity reporting or you know standard setting, it is a worthy investment if we are going to like see a upcoming existential threats six years in the future.
…And to me that means trusting the people, because you’re yielding your most important power, which is agenda setting power to the people.
謝謝行政院及vTaiwan給我們機會參加今天的討論。因為我不是台灣人，所以發言不清楚的地方，請大家見諒。台灣政府對於共享經濟的相關議題採取如此開放立場，我們非常感激，透過今天的簡報我們向大家、經濟部說明Airbnb商業模式及我們與各國政府一同合作針對分享住宿新創法規及分享經濟的影響。 事實上這個圖片是指我們公司如何開始的，其實我覺得這個故事是比較重要，因為代表著我們在這裡工作的精神及使命是什麼。 Airbnb是在2008年開始，如同許多新創的公司一樣，我們的創辦人為了解決一個問題，他的問題就是為了補貼在舊金山的房租，他們看到自己的住處有不用的空間，他們想要把airspace放到網上，剛好這個時間有一個大型的design conference在San Francisco，很多的旅客許多住宿，飯店也滿了，有三位房客在網上找到air bed and breakfast，就是剛剛我們的presentation有講的，現在都叫「Airbnb」。 Airbnb現在191個國家、3.4萬城市、共有150萬多個房源，這是我們比較新的數據，剛剛去拿的。其中最近加入的一個國家是古巴，如今Airbnb已經成為一個全球的平台，經營在一個多元化的住宿市場，包括傳統民宿，其實我們最重視的是跟創辦人一樣，願意在平台上分享閒置的空間，分享給全世界的旅客。 其實住宿分享並不是新的一件事，其實是最原始的旅行方式，但因為現在的網路及科技可以讓用戶安全及簡單預定體驗當地人的生活，當初創辦人接待頭三名的房客到現在，Airbnb現在已經在全球接待超過5000萬的房客，在這個夏天，Airbnb達到最高峰的一天，我們一個晚上在全世界有100萬的Airbnb房客住在我們的房源裡面。 交通部剛剛presentation也有介紹過我們的運作怎麼樣，我們可以快速增長是因為我們有訓練，為了有效訓練，我們做了很多不同的措施及努力。最基本的就是旅行結束的時候，我們會提供一個雙向的評價系統，這個系統是公開的，可以讓未來的用戶參考，可以確保房東提供最好的接待服務，及房客可以遵照房東的平台；我們也會決定房客入住24小時後才將付款轉給房東，這是給我們房客的一些保障。為什麼我們會這樣做？如果房東跟房客在24小時裡面沒有問題的話，這是技術上的continue；但是如果有問題的話，Airbnb會先保留付款，直到我們把房東跟房客間的問題解決。 其他保護措施還有：24小時顧客熱線，我們在亞太區operations hold現在也有中文的協助。 我們也有以250人組成的信任小組，叫「信任與安全小組」（ trust and safety），在全球持續保護我們的消費者。在今年夏天Airbnb總共有1400萬人在Airbnb旅行，我們接到的緊急的電話僅300通…
…And because this is the collective will of the stakeholders, when we implement that, it also increase trust .
…AI strategy, moving beyond traditional defensive measures like chip export controls to a more proactive approach focused on bolstering domestic research and development and accelerating its leadership in the AI domain. I attended the Trusted Tech Summit in Washington in early December and felt the momentum behind this shift firsthand. Experts at the summit generally agreed that relying on a purely defensive strategy will not sustain America’s long-term advantages. Against this backdrop, an “proactive strategy” is emerging, one that involves collaborating with international allies to jointly develop AI systems that can assist…
