2026 INAHTA 大會 Day 1：AI 與醫療科技評估的交匯點

今年有幸虛擬參與了 2026 年 INAHTA（The International Network of Agencies for Health Technology Assessment，國際官方醫療科技評估組織聯盟）在土耳其伊斯坦堡舉辦的年度大會。本年度大會的核心主軸緊扣時代脈動，聚焦於「人工智慧（AI）與全生命週期醫療科技評估（Lifecycle HTA）的挑戰與機遇」。大會採取了嚴格的保密與不對外引述原則，為全球與會者營造了一個充滿信任與安全感的專屬交流空間，使大家得以暢所欲言，聽見各國機構面對 AI 狂潮時最真實的痛點、掙扎與突破。

觀察一：從「試水溫」到「務實應用」——AI 在 HTA 的定位轉變

國際態度的演變

國際間對 AI 的態度已經從初期的「試水溫」與觀望，逐漸轉向「務實且有策略的應用」。多個國際合作組織的調查與經驗分享皆顯示，目前 AI 在 HTA 領域中最具潛力的應用場景，並非用來取代專業的醫療與經濟決策，而是接手繁瑣的行政庶務與資料處理工作。

具體應用案例

許多機構開始運用 AI 協助文獻篩選、會議紀錄摘要，或是將艱澀的臨床指引轉化為淺顯易懂的民眾版衛教素材。會中有一段極為生動的比喻令人印象深刻：

AI 就像是一名急於給出正確答案以討好教授的學生，或是擅長執行特定技術的工匠

它能迅速找出資料庫中的模式，但我們絕不能將「價值判斷」與「最終決策」交給它。

AI 應用的核心目的

將枯燥的工作交接給 AI，真正的目的是為了釋放 HTA 專家們寶貴且不可思議的大腦，讓大家能專注於更核心的深度分析與評估。

觀察二：資料安全與 AI 素養之間的平衡——從組織到個人

面臨的雙重挑戰

資料安全與「個人 AI 素養（AI Literacy）」之間的拉扯，是目前各國共同面臨的巨大挑戰。在 HTA 領域中，我們經常需要處理尚未公開的試驗數據或具極高機密性的廠商資料。

組織層級的防護措施

為了解決資料外流的隱憂，許多機構正積極建構封閉的安全網路，甚至推動「只共享知識，不共享數據」的聯合知識系統。

最大的風險：人為因素

然而，現場的討論也尖銳地指出，這其中最大的風險其實是「人」。即使組織建立了再嚴密的安全架構與內部 AI 系統，如果員工個人缺乏足夠的 AI 識讀能力與資安意識，在日常工作中不慎將機密資訊輸入至公開的外部 AI 模型中，所有的防線都將形同虛設。

由下而上的必要性

因此，建立組織政策固然重要，但由下而上地提升每一位工作人員的 AI 意識與操作素養，更是導入 AI 工具前不可或缺的基礎工程。

觀察三：跨國合作——從理想口號到生存策略

技術更迭的高不確定性

面對 AI 技術日新月異的發展，尋求「跨國合作」已從一個理想的口號，變成無可迴避的生存策略。許多與會代表坦言，由於技術更迭太快，我們幾乎無法在當下定義出一套永恆不變的「最佳實踐（Best Practice）」，因為當規範寫好時，工具可能又已經升級了。

避免重複發明輪子

面對這樣的高不確定性，與會者們達成了一個強烈的共識：我們必須避免「重複發明輪子」。與其讓各國機構各自花費數月時間去測試與評估同一套 AI 工具的優劣，不如透過國際聯盟的力量，建立透明的經驗交流平台與動態的治理框架。

深度協作的迫切性

在這個充滿顛覆性變革的時代，我們比以往任何時候都更迫切地需要彼此的緊密協作與資源共享。

展望：在變革中堅守 HTA 的嚴謹與信任

第一天的議程在熱烈且坦誠的對話中畫下完美的句點，也為接下來更深入的世界咖啡館討論奠定了極佳的基礎。科技的進步無可阻擋，而我們如何在這波浪潮中堅守 HTA 的嚴謹與信任，將是我們共同的課題。

大家在各自的專業領域中，又是如何平衡 AI 應用與機密資料安全的呢？歡迎在底下留言交流或分享！

English version below

INAHTA 2026 Istanbul Conference Day 1: The Intersection of AI and Health Technology Assessment

I was honored to participate virtually in the 2026 INAHTA (The International Network of Agencies for Health Technology Assessment) annual conference held in Istanbul, Turkey. This year’s conference centered on a timely theme: “Challenges and Opportunities in Artificial Intelligence and Lifecycle Health Technology Assessment (HTA).” The conference maintained strict confidentiality and non-attribution principles, creating a trusted and secure space for global participants to openly share their most authentic pain points, struggles, and breakthroughs as they face the AI wave. The first day’s agenda was filled with profound discussions.

Observation One: From “Testing Waters” to “Pragmatic Application”—AI’s Evolving Role in HTA

The Shift in International Perspective

The international attitude toward AI has gradually shifted from initial “exploration and observation” to “pragmatic and strategic application.” Surveys and experience sharing from multiple international collaborative organizations reveal that the most promising applications of AI in the HTA field are not to replace professional medical and economic decision-making, but rather to handle tedious administrative tasks and data processing work.

Concrete Implementation Examples

Many institutions are beginning to use AI to assist with literature screening, meeting minute summarization, or converting complex clinical guidelines into easy-to-understand public health education materials. One particularly vivid analogy from the conference left a lasting impression:

AI is like an eager student wanting to give the right answer to impress the professor, or a craftsman skilled at executing specific techniques.

It can quickly identify patterns in databases, but we must never entrust “value judgment” and “final decision-making” to it.

The True Purpose of AI Application

The real goal of delegating tedious work to AI is to liberate the precious and remarkable minds of HTA experts, enabling them to focus on more core, in-depth analysis and assessment.

Observation Two: Balancing Data Security and AI Literacy—From Organization to Individual

The Dual Challenge

The tension between data security and “personal AI literacy” is a major challenge currently facing all countries. In the HTA field, we frequently handle unpublished trial data or highly confidential proprietary information.

Organizational-Level Safeguards

To address data breach concerns, many institutions are actively building secure closed networks and even promoting “knowledge-sharing only, not data-sharing” collaborative knowledge systems.

The Greatest Risk: The Human Factor

However, discussions at the conference also pointed out sharply that the greatest risk is actually “people.” No matter how robust an organization’s security infrastructure and internal AI systems are, if employees lack sufficient AI literacy and cybersecurity awareness and inadvertently input confidential information into public external AI models during daily work, all defensive lines become useless.

The Necessity of Bottom-Up Efforts

Therefore, while establishing organizational policies is important, building AI awareness and operational literacy among every worker from the bottom up is an essential prerequisite before introducing AI tools.

Observation Three: International Collaboration—From Ideal Slogan to Survival Strategy

The High Uncertainty of Rapid Technological Change

Faced with ever-evolving AI technology, seeking “international collaboration” has transformed from an ideal motto into an unavoidable survival strategy. Many conference representatives candidly acknowledged that due to rapid technological iteration, we can hardly define a set of eternally unchanging “best practices” today, because by the time regulations are written, the tools may have already upgraded.

Avoiding Reinventing the Wheel

Faced with such high uncertainty, conference attendees reached a strong consensus: we must avoid “reinventing the wheel.” Rather than having each country’s institution spend months testing and evaluating the same AI tool individually, it is better to leverage the power of international alliances to establish transparent experience-sharing platforms and dynamic governance frameworks.

The Urgency of Deep Collaboration

In this era of disruptive change, we need each other’s close collaboration and resource sharing more urgently than ever before.

Outlook: Maintaining HTA’s Rigor and Trust Amid Change

The first day’s agenda concluded with heated and candid dialogue, laying an excellent foundation for more in-depth world café discussions to follow. Technological progress is inevitable, and how we maintain HTA’s rigor and trust in this wave of change is our shared challenge.

How do you balance AI applications with confidential data security in your respective professional fields? I welcome your comments and insights below!