The Committee appreciates the government’s ongoing efforts to address the evolving challenges of the digital economy. However, several cross-cutting issues now require more urgent and coordinated action. In particular, the absence of a unified legal framework governing online content moderation has led to overlapping and inconsistent guidance across agencies. This fragmentation has created significant uncertainty for digital platforms, resulting in over-censorship and unintended constraints on democratic discourse and innovation. At the same time, the growing number of government requests for user data without clear legal safeguards or transparent processes raises serious concerns regarding privacy, due process, and human rights. Without urgent regulatory modernization, Taiwan risks regulatory fragmentation that threatens both its democratic credibility and economic competitiveness in the digital age.
To ensure Taiwan remains competitive in the global digital economy, the Committee urges the government to establish a clear and harmonized legal framework for online content governance, including transparent standards for data requests and user rights protections. In parallel, the government should accelerate policy development and financial incentives to support edge AI deployment, particularly by empowering local developers and SMEs to participate in next-generation AI innovation. Finally, the creation of a universal, cross-sectoral regulatory sandbox is essential to enabling agile policy experimentation and promoting responsible innovation across emerging technologies. Taken together, these actions will help preserve Taiwan’s democratic values while strengthening its position as a global digital leader.
Suggestion 1: Establish clear intermediary liability frameworks to strengthen Taiwan’s digital economy.
The Committee is concerned that inconsistent regulatory approaches to online content are undermining Taiwan’s digital development. Despite past recommendations for clear safe harbor principles, no meaningful progress has been made. In fact, the trend toward regulatory fragmentation appears to have continued, further complicating the operating environment. Agencies continue to issue disparate rules and require proactive content monitoring, creating legal uncertainty and undue burdens.
This lack of clarity hinders digital platforms’ ability to operate effectively in Taiwan and poses risks to democratic principles such as freedom of expression, due process, and participatory digital governance. It also dampens innovation, investment, and growth in Taiwan’s digital economy. The Committee urges the Executive Yuan to establish comprehensive and harmonized intermediary liability guidelines that align with international standards. These should balance the protection of online safety with the preservation of free expression and support for economic development.
A notable example of regulatory overreach is the Ministry of Health and Welfare’s (MOHW) enforcement of the Tobacco Hazards Prevention Act. The MOHW has adopted an expansive definition of “advertising” to include user-generated content, imposed penalties despite platforms’ compliance efforts, and required keyword filtering that amounts to pre-screening. Of particular concern is the MOHW’s proposed legislative amendment that would mandate proactive monitoring of online content, an approach that conflicts with global best practices and undermines Taiwan’s commitment to internet freedom. These developments have occurred without adequate cross-agency coordination or engagement with affected stakeholders, raising concerns about regulatory transparency and fairness.
To promote a healthy and innovation-friendly digital ecosystem, the Committee reiterates the importance of adopting clear safe harbor provisions specified below. By aligning its regulatory framework with democratic norms and global best practices, such as those outlined in the Manila Principles, Taiwan can safeguard free expression, foster innovation, and enhance its international standing as a leader in digital rights and governance.
- Establish overarching guidelines. The Committee calls on the National Communications Commission (NCC) and the Ministry of Digital Affairs (MODA) to develop, under the Executive Yuan’s guidance, cross-agency intermediary liability and safe harbor rules based on international norms like the Manila Principles. The rules can be established as formal administrative guidelines or regulations, intended to act as the immediate interpretive standard for existing laws. Furthermore, they should inform and guide any targeted legislative proposals should such become necessary.
- Adhere to core principles. Ensure that the developed rules, which would apply to all relevant laws regulating online content, are based on the following core principles:
a.) Encourage Good Samaritan actions. Platforms that conduct reasonable, voluntary content moderation should be protected.
b.) Promote “notice-and-takedown” practices. This mechanism should serve as the primary method for handling illegal user-generated content (UGC), with clear notice rules and prompt platform action.
c.) Distinguish commercial content from UGC. UGC generally should not be pre-screened but should instead be taken down after due notice. Such an approach allows enforcement efforts to focus on illegal commercial activity while still safeguarding user expression.
d.) Oppose general monitoring obligations. Universal UGC monitoring or pre-screening is impractical and greatly hurts free speech and communication secrecy. Further, universal UGC monitoring damages Taiwan’s democratic standing, contravenes international practice, and creates non-tariff barriers for cross-border services.
e.) Acknowledge service diversity. Rules should be tailored to different types of internet services, such as search and indexing, hosting, and interpersonal communication, to avoid inflexible, one-size-fits-all approaches.
f.) Ensure government transparency. Due process, transparency, and accountability should be required when government agencies request content removal or user data.
g.) Draw on prior frameworks. Valuable principles from the NCC’s draft Digital Intermediary Services Act, such as safe harbor and Good Samaritan clauses, can serve as illustrative examples for future policymaking. - Implement central oversight for cross-ministry adherence. An Executive Yuan-led monitoring mechanism, facilitated by the NCC and MODA, should be established to actively ensure that all relevant government ministries consistently apply overarching intermediary liability guidelines when developing new regulations and enforcing laws concerning online platforms. A clear, principles-based, and harmonized intermediary liability framework led by the Executive Yuan is essential for giving businesses the legal certainty they need to take effective action against illegal content while safeguarding free expression.
Suggestion 2: Establish clear guidelines and safeguards for government requests for user data.
The Committee stresses the challenges faced by digital service businesses operating in Taiwan due to increasing and unfettered government requests for user data in the name of administrative investigation. The lack of a clear and precisely defined legal framework, coupled with inadequate understanding of technical capabilities and potential conflicts with international laws, creates uncertainty for international digital service providers.
This growing trend of government surveillance undermines trust in the digital economy and raises serious concerns about human rights and individual freedoms. The absence of due process and rights to appeal by both data subjects and digital services providers exacerbates these concerns.
- Develop and publish a clear and comprehensive regulatory framework. A clear regulatory framework should be established to manage government requests for user data, outlining specific procedures and criteria for such requests. This framework should protect the right to privacy and self-determination over personal data. Current laws, like the Personal Data Protection Act and provisions scattered across different administrative laws, are insufficient. The framework should adopt international best practices, such as requiring valid legal processes (including court orders for certain types of data disclosure), scrutinizing requests for legality and scope, and challenging overly broad or unclear requests, as seen in the Stored Communications Act and relevant global data protection regulations.
- Provide training and education for government officials in the technical capabilities and limitations of digital service providers. Key aspects to be included are the need to respect user privacy, an introduction to international best practices, and the fact that service providers may not possess certain data due to user deletion or retention policies.
- Establish a centralized monitoring system to monitor government requests for user data. This system should ensure that requests are documented, reviewed, and approved by an independent authority. In the interest of transparency, it should also facilitate the issuance of regular reports detailing these requests by type and how they were handled.
- Ensure formal and transparent request channels. Government requests for user data should be made through formal and transparent channels, using legally valid templates and citing specific laws and regulations, as well as the legal grounds for making such requests. This approach would be in line with international standards such as the U.S. Electronic Communications Privacy Act (ECPA) and Japan’s Guidelines for Protection of Personal Information in Telecommunications Business.
- Introduce robust due process protections for data subjects and digital services providers, including the right to appeal and challenge government requests. Users should be notified about legal demands when appropriate, allowing them to seek legal recourse. Doing so ensures that potential misuse or overreach can be challenged, maintaining legal process integrity and protecting individual rights.
- Eliminate the mandatory data retention period. Requirements that service providers retain data for a certain period of time for purposes of administrative investigation should be eliminated in existing laws and future regulations to prevent over-collection of user data and lower the risk of data breaches. By addressing these concerns and establishing a clear and predictable system for government requests for user data, Taiwan can promote a safe and trustworthy digital environment while also protecting the rights and interests of both businesses and individual users. Immediate action in this regard would help ensure that Taiwan remains a welcoming destination for investment.
Suggestion 3: Prioritize edge AI and provide incentives to developers and SMEs to accelerate AI deployment.
The Committee welcomes the government’s commitment to making Taiwan a global leader in artificial intelligence and preparing industries and society for widespread AI integration. As technological innovation drives the development of smaller, more efficient models, on-device AI inference (known as edge AI) has become increasingly viable. Unlike cloud-based AI, edge AI processes data locally on devices, improving speed and privacy and enabling scalable, cost-effective deployment across a wide array of devices, including smartphones, PCs, IoT equipment, XR platforms, and automotive applications. Edge AI offers significant advantages in latency reduction, data privacy, operational efficiency, and cost savings.
Moreover, edge AI presents unique opportunities for small and medium-sized enterprises (SMEs), which often lack the financial or technical resources to build and train large-scale models. With its globally competitive ICT industry and mature AI and PC ecosystems, Taiwan is ideally positioned to lead in the adoption and development of edge AI. To fully realize this potential, the Committee encourages the government to provide targeted policy support and incentives aimed at cultivating a developer ecosystem and accelerating industry-wide deployment, particularly for SMEs.
The Committee commends the AI talent cultivation and industry development programs initiated by the Ministry of Economic Affairs (MOEA) and the Ministry of Digital Affairs (MODA), which have laid a strong foundation for AI readiness across sectors. To sustain and expand these efforts, we respectfully offer the following policy recommendations:
- Integrate hybrid and edge AI into government AI training programs. Both the MOEA and MODA are actively advancing AI ecosystem development through training courses and community-building initiatives. We recommend that these programs be further strengthened by:
- Consolidating cross-ministerial AI training plans under a centralized platform or task force led by the Executive Yuan to ensure coherence and efficiency.
- Incorporating dedicated modules on hybrid and edge AI, with a focus on the practical application of small-scale, high-performance models for on-device inference.
- Partnering with international technology leaders to make advanced developer tools and up-to-date training resources widely accessible.
- Regularly updating training content to reflect the evolving AI landscape, including sector-specific use cases that demonstrate the economic and operational benefits of edge deployment.
- Provide targeted incentives to promote AI adoption among SMEs. To support broader AI integration, particularly among smaller enterprises, the Committee encourages the introduction of tailored incentive mechanisms, such as:
-
- Financial support (including subsidies, matching grants, or tax deductions) for AI hardware and software acquisition, development, and deployment.
- Non-financial support (including technical mentorship, developer matchmaking, regulatory guidance, and pilot program participation) to reduce adoption barriers.
- Sector-based incentive programs that align with national economic priorities and drive AI innovation in key verticals, such as manufacturing, logistics, healthcare, and education.
As AI deployment continues to evolve, driven by lower training costs, faster inference, and an increasing emphasis on edge computing, the Committee urges the government to adapt its strategy to ensure that Taiwan’s digital economy remains agile, inclusive, and competitive. By enabling SMEs and developers to access the tools, knowledge, and support they need, Taiwan can unlock the full potential of edge AI and establish itself as a global leader in next-generation AI applications.
Suggestion 4: Establish a universal regulatory sandbox for innovation in Taiwan.
Taiwan’s long-standing leadership in technology and innovation has positioned it as a critical hub in the global digital economy. However, the absence of cross-sectoral mechanisms for policy experimentation poses growing challenges for businesses attempting to deploy emerging technologies. While targeted regulatory sandboxes in areas such as fintech and autonomous vehicles have delivered tangible benefits, no comparable frameworks currently exist for other high-potential sectors, including healthcare, tourism, transportation, and smart cities.
Without structured pathways for real-world testing and regulatory engagement, businesses face persistent legal uncertainty. This inhibits investment, discourages digital transformation, and limits Taiwan’s ability to remain competitive in fast-evolving sectors, particularly as the global adoption of AI and other next-generation technologies accelerates.
To address this gap, the Committee recommends the establishment of a universal regulatory sandbox that enables controlled experimentation across all industries. This initiative would provide a transparent, time-bound environment in which companies can pilot novel technologies while regulators evaluate their risks and benefits. A well-designed sandbox should balance policy flexibility with public interest protections, enabling innovation to proceed safely and effectively.
- Appoint the National Development Council (NDC) as the lead coordinating agency. To ensure cross-ministerial alignment, the NDC should serve as the central coordinating body for the universal sandbox framework. This role would help facilitate collaboration among regulatory agencies and minimize the risk of fragmented or contradictory policies across different sectors. The Committee further recommends that the NDC serve as a central point of contact for foreign-invested enterprises seeking participation, reinforcing Taiwan’s openness to international innovation.
- Develop a clear, business-friendly testing framework. The sandbox should offer defined timelines, transparent eligibility criteria, and clearly articulated conditions under which participating firms may test their products or services. The framework should be adaptable to varying risk profiles and technology maturity levels while ensuring that regulatory requirements are proportionate and innovation-enabling.
- Ensure predictable pathways to full market entry. Technologies or solutions that demonstrate societal benefit and operational feasibility during the testing phase should be granted an efficient and predictable route toward formal regulatory approval. The transition out of the sandbox must be structured to reward responsible innovation and minimize regulatory uncertainty following pilot completion.
- Use sandbox outcomes to inform broader regulatory reform. Lessons learned from sandbox trials should be systematically incorporated into broader policymaking processes. This data-driven approach will help modernize Taiwan’s regulatory environment in a way that reflects real-world applications rather than theoretical concerns, aligning oversight mechanisms with emerging trends in AI, data use, digital platforms, and sustainability.
A universal regulatory sandbox would not only strengthen Taiwan’s global competitiveness but also reinforce its reputation as a forward-looking, innovation-driven economy. The Committee stands ready to support the government in designing and operationalizing this framework to ensure it delivers tangible outcomes for businesses, regulators, and society at large.
本委員會肯定臺灣政府持續處理數位經濟所帶來的多變挑戰。然而,仍有數項亟需加速推動跨部會協調與統一應對的關鍵議題。
尤其是,缺乏統一的網路內容管理的法律框架,導致各主管機關所發布的指導方針彼此不一致且迭有重疊。這種碎片化的監管對數位平台造成顯著的不確定性,連帶導致內容的過度審查,並對民主論述與創新加諸非預期的限制。此外,政府在缺乏明確法律保障或透明程序的情況下,對用戶資料的調閱要求日益增加,引發本委員會對隱私權、正當法律程序及人權的嚴正關切。倘若未能及時完成法規現代化,臺灣恐將面臨監管持續碎片化的風險,進而危及其在數位時代的民主公信力與經濟競爭力。
為確保臺灣在全球數位經濟中維持競爭力,本委員會建議政府建立一套明確且協調一致的網路內容治理法律框架,並納入資料調取之公開標準並保障使用者權利。同時,政府亦應加速制定與邊緣AI(Edge AI)相關政策與獎勵措施之規劃與推動,特別應賦能本地開發者與中小企業參與下一代AI的創新。最後,建立一個普遍適用、跨領域的監理沙盒(Regulatory Sandbox),將有助於推動敏捷的政策實驗,並促進新興技術領域中負責任的創新。總而言之,上述各項行動將有助於維護臺灣的民主價值,同時強化臺灣作為全球數位領導者的地位。
建議一:建立明確的網路中介者責任框架,強化臺灣數位經濟發展
委員會擔憂,目前臺灣對於網路內容不一致的監管方式,正在損害臺灣的數位發展。儘管過去曾建議建立明確的避風港原則,但並未達到實質進展。事實上,監管碎片化的趨勢仍在持續,進一步加劇營運環境的複雜性。各主管機關不斷發布不一致的規定,並要求平台主動監控內容,造成了法律上的不確定性與過度負擔。
這種不明確性阻礙了數位平台在臺灣的有效營運,並對言論自由、正當法律程序及參與式數位治理等民主原則構成風險。同時也抑制了臺灣數位經濟的創新、投資與成長。本委員會強烈建議,行政院建立一套全面且一致的網路中介者責任指導方針,並應參照國際標準。相關規範應兼顧線上安全保障與言論自由維護,並支持經濟發展。
其中一個明顯過度監管例子是衛生福利部(下稱衛福部)對《菸害防制法》的執法方式。衛福部將「廣告」一詞採取過度擴張的定義,並將使用者生成的內容(UGC)也納入其中,即使平台已盡力配合仍被裁罰,同時要求進行網路關鍵字過濾篩選,實質上形同預先審查使用者的內容。尤其令人擔憂的是,衛福部研擬的修法草案要求平台主動監控線上內容,此舉不僅違背國際最佳實務,更嚴重損害臺灣對網路自由的承諾。上述措施缺乏充分的跨部會協調,且亦未與受影響的利害關係人溝通,引發本委員會對監管透明度及公平性的高度關切。
為促進健康且有利於創新的數位生態系,本委員會重申採納明確避風港條款的重要性。臺灣必須透過確保其監管框架符合民主常理,並與《馬尼拉原則》等全球最佳實踐接軌,藉此保障言論自由,促進創新,並提升其在數位人權和網路治理的國際領導地位。
- 制定全面指導方針
委員會呼籲國家通訊傳播委員會及數位發展部,應在行政院的指導下,依據《馬尼拉原則》等國際規範,共同制定跨部會的網路中介者責任及避風港條款。相關條款可透過正式行政指導方針或法規形式訂定,作為現行法律的優先解釋依據。此外,若未來需要提出相關立法提案時,此等條款亦應提供參考依據與政策指引。 - 遵循核心原則
應確保制定的條款,適用於任何涉及網路內容規範的相關法律,並且符合以下核心原則。
a.) 鼓勵「善良的撒瑪利亞人(Good Samaritan)」行為:對進行合理且自願內容管理的平台,應予以保護。
b.) 推廣「通知後下架」(notice-and-takedown) 機制:此機制應作為處理非法使用者生成內容(UGC)的主要途徑,並應明確訂定通知程序及平台處置方式。
c.) 明確區分商業廣告內容與使用者生成內容(UGC):使用者生成內容(UGC)不應預先審查,而應在收到適當通知後再行移除。此舉能使主管機關將執法重點聚焦於非法商業行為上,同時也保護使用者的言論自由。
d.) 反對全面監控義務:委員會強烈反對要求平台針對使用者生成內容(UGC)施加全面監控或預先審查;此做法既不切實際,也嚴重侵害言論自由和通訊秘密。再者,對使用者生成內容全面監控亦損及臺灣的民主形象,違反國際慣例,並構成跨境服務的非關稅貿易壁壘。
e.) 承認網路服務的多樣性:相關規範應根據不同類型的網際網路服務加以制定,例如:搜尋及索引、網站託管服務、個人通訊等,以避免僵化、一體適用的制定方式。
f.) 確保政府程序之透明:政府機關要求平台移除內容或調取使用者資料時,應符合正當法律程序、資訊透明度及問責原則。
g.) 借鑒現有政策原則:國家通訊傳播委員會(NCC)所提出的數位中介服務法草案中,關於避風港條款及「善良的撒瑪利亞人」條款等原則,可作為未來政策制定的重要參考。 - 建立一套跨部會一致性適用的中央監督機制:
應建立一套由行政院主導,並由國家通訊傳播委員會及數位發展部協助推動的監督機制,此機制旨在積極促使所有相關政府部會,在研擬、修訂與執行與線上平台相關的法規與法律時,遵循一致性的網路中介者責任指導方針。建立這樣一套由行政院主導、清晰、基於原則且協調一致的網路中介者責任框架,是提供企業必要法律確定性的關鍵,從而使得企業能在有效打擊非法內容的同時,也能保障使用者的言論自由。
建議二:針對政府對使用者資料的請求制定明確的指導方針和保障措施
委員會強調,由於政府以行政調查的名義,進行越來越多且不受節制的用戶個資調取,對在臺灣營運的數位服務企業面臨挑戰。這因為缺乏明確和精確定義的法律框架,加上對技術能力的了解不足,以及與國際法的潛在衝突,在對國際數位服務提供者帶來了不確定性。
日益增長的政府監控之趨勢,破壞了人們對數位經濟的信任,並引發了對人權和個人自由的嚴重顧慮。對個資主體和數位服務提供者,由於缺乏正當程序的保護和可救濟的權利,均加劇了這些擔憂。
建議:
- 制定並發佈清晰、全面的法規架構。 應建立明確的法規框架來管理監督政府對使用者個人資料的請求,勾勒此類請求應具有的具體程序和標準。該框架應保護個人資料的隱私權和自決權。現行法律,如《個人資料保護法》和分散在不同行政法中的規定不足的。該框架應採用國際最佳實踐,例如要求有效的法律程序(包括針對某些類型的資料揭露的法院命令),審查合法性和範圍的請求,並對過於寬泛或不明確的請求提出質疑,正如美國聯邦《儲存通訊記錄法》和相關的國際資料保護法規所揭示的。
- 為政府官員提供有關數位服務提供者的技術能力和局限性的培訓和教育。尤其是需要尊重用戶隱私、對於國際最佳實踐的介紹,以及服務提供者可能由於用戶刪除或保留政策而無法擁有某些資料。
- 建立集中監督系統來監督政府對使用者資料的請求。該系統應確保請求由獨立機構記錄、審查和批准。為了提高透明度,還應促進定期報告的發布,按類型詳細說明這些政府請求及其處理方式。
- 確保請求管道正式、透明。 政府對使用者資料的請求應透過正式和透明的管道提出,使用合法有效的範本並引用具體的法律法規以及提出此類請求的法律依據。此方法將符合美國《電子通訊隱私法》(ECPA) 和日本《電信業務個人資訊保護指南》等國際標準。
- 為資料主體和數位服務提供者引入強而有力的正當程序保護,包括上訴和質疑政府要求的權利。用戶應在適當的時候被通知有關法律要求其個資之情形,以便他們尋求法律救濟。這確保了對潛在的濫用或越權行為可提出挑戰,從而維護法律程序的完整性並保護個人權利。
- 取消強制性資料保留期。要求現有法律和未來的法規應取消服務提供者出於行政調查目的而在一定期限內保留資料的做法,以防止過度收集用戶資料並降低資料外洩的風險。透過解決這些問題並針對政府對使用者資料的請求建立明確且可預測的系統,臺灣可以促進安全可靠的數位環境,同時保護企業和個人使用者的權益。在這方面立即採取行動將有助於確保臺灣仍然有受歡迎的投資環境。
建議三:優先考慮邊緣AI,並為開發者和中小企業提供獎勵措施,以加速AI部署
委員會歡迎政府致力於使臺灣成為全球人工智慧(AI)領導者,並為產業和社會做好廣泛AI整合的準備。隨著技術創新推動更小、更高效模型的發展,設備上的AI推論(即邊緣AI)變得越來越可行。與基於雲端的AI不同,邊緣AI在設備上本地處理數據,提高了速度和隱私,並能夠在包括智能手機、個人電腦、物聯網設備、XR平台和汽車應用在內的各種設備上進行可擴展、成本效益高的部署。邊緣AI在降低延遲、數據隱私、運營效率和成本節省方面具有顯著優勢。
此外,邊緣AI為中小企業提供了獨特的機會,這些企業通常缺乏建立和訓練大規模模型的財務或技術資源。憑藉其全球競爭力的ICT產業和成熟的AI和PC生態系統,臺灣在邊緣AI的採用和發展方面處於理想位置。為了充分實現這一潛力,委員會鼓勵政府對此提供特定的政策支持和獎勵措施,以培育開發者生態系統並加速百工百業的部署,特別是針對中小企業。
委員會讚揚經濟部和數位發展部啟動的AI人才培育和產業發展計劃,這些計劃為各行各業的AI準備工作奠定了堅實的基礎。為了維持和擴展這些努力,我們謹提出以下政策建議:
- 將混合AI和邊緣AI整合到政府的AI培訓計劃中:經濟部和數位發展部都在透過培訓課程和社群計劃積極推動AI生態系統的發展。我們建議進一步加強這些計劃,具體措施包括:
.在行政院領導下,將跨部門的AI培訓計劃整合到一個集中平台或工作小組中,以確保連貫性和效率。
.包含混合AI和邊緣AI的課程模組,重點關注小規模、高性能模型的實際應用。
.與國際技術領導者合作,使先進的開發工具和最新的培訓資源廣泛可用。
.定期更新培訓內容,以反映不斷發展的AI領域,包括展示各產業部署邊緣AI應用的經濟和營運效益。 - 提供特定的獎勵措施以促進中小企業的AI採用:為了支持更廣泛的AI整合,特別是針對較小的企業,委員會鼓勵引入量身定制的獎勵機制,例如:
.財務支持(包括補貼、配套資助或稅收減免)以取得、開發和部署AI硬體和軟體。
.非財務支持(包括技術指導、開發者媒合、法規指導和試點計劃參與)以減少採用障礙。
.與國家經濟優先事項一致的產業獎勵計劃,以推動如製造、物流、醫療保健和教育等關鍵領域的AI創新。
隨著AI部署的不斷發展,受益於更低的訓練成本、更快的推論速度和對邊緣運算的日益重視,委員會敦促政府調整其戰略,以確保臺灣的數位經濟保持靈活、包容和競爭力。經由使中小企業和開發者能夠獲得所需的工具、知識和支持,臺灣可以釋放邊緣AI的全部潛力,並確立自己在下一代AI應用中的全球領導地位。
建議四:建立標準化且支持創新的監理沙盒機制
長期以來,臺灣在科技與創新領域具有領先地位,並已成為全球數位經濟體系的關鍵樞紐。然而,目前尚缺乏跨部會、跨產業的政策試驗機制,對有意導入新興科技的企業而言,形成日益嚴峻的挑戰。儘管政府已於金融科技、自駕車等特定領域建立監理沙盒機制,並已取得具體成效,但在包括醫療、觀光、交通與智慧城市等具高度發展潛力的產業,至今仍無類似的政策試驗框架。
在缺乏制度化的實務驗證與政策對話管道下,企業長期面臨法規不確定性,導致投資意願受抑、數位轉型停滯,也削弱臺灣在快速迭代的產業中的國際競爭力。在全球加速導入人工智慧與其他創新技術之際,此一問題將更為顯著。
為填補此一法規制度缺口,委員會呼籲政府推動一套標準化的創新監理沙盒機制,使各產業皆能在可控條件下進行技術試驗。該機制將提供一個具透明度與時限性的制度環境,使企業得以導入與測試創新技術,並讓主管機關據以評估其潛在風險與社會效益。完善的沙盒機制設計應平衡政策彈性與公共利益,確保科技創新能在安全、有序的前提下發展。
為支持產業創新與成長,以促進投資為導向的沙盒機制應奠基於以下四項核心原則:
- 由國家發展委員會(下稱國發會)擔任統籌協調角色
為確保跨部會協作,委員會建議應由國發會擔任統籌窗口,協助各主管機關整合資源,避免產業間適用標準與程序產生不一致或重疊。此外,委員會亦建議由國發會擔任外資企業申請參與沙盒試驗之聯絡窗口。有鑒於外商時有遭遇主管機關以「涉政治敏感性」為由而反覆或遲疑回應之情事,若能於國發會之層級提供穩定且具一致性的協調角色,有助於彰顯臺灣對國際創新的開放立場。 - 建立明確、投資友善的試驗框架
沙盒機制應明確規範適用期間、申請資格與適用條件,讓企業得以在可預期的制度下測試其產品或服務。同時,該框架亦應具備彈性,依據不同風險屬性與技術成熟度進行調整,同時確保相關法規要求符合比例原則,並能實質支持推動創新。 - 建立可預期的市場導入機制
對於在試驗階段展現公眾利益與技術可行性的創新方案,政府應提供明確且高效率的法規銜接路徑,使其能順利取得正式核可並投入市場。沙盒試驗退場機制應具制度化設計,獎勵負責任的的創新,同時降低企業常於試驗結束後面臨的法規不確定性。過往已有多起案例顯示,企業即使成功完成沙盒試驗,仍常面臨政治干預或監管延宕,導致創新無法真正落地。有效的沙盒制度,不應只是提供試驗場域,更應成為支持創新成果擴大應用的制度通道。 - 沙盒試驗成果應成為推動整體法規改革的重要依據
試驗過程中累積的數據與經驗,應被系統性地納入後續政策制定與制度調整流程,作為推動法規改革的重要依據。政府應擺脫僅憑理論預設進行規範設計的舊有模式,轉向以數據為基礎、以實證為依據的改革方向,以更貼近實務應用情境。唯有如此,臺灣的監理體系才能與人工智慧、資料應用、數位平台及永續發展等新興趨勢同步演進,真正建立具前瞻性與韌性的政策架構。
建立一套標準化且支持創新的跨領域監理沙盒機制,不僅能強化台灣在全球市場的競爭力,更鞏固其作為一個具備前瞻視野與創新實力的國際經濟體之地位。委員會願與政府密切合作,協助推動制度設計與實務執行,確保該機制能在促進產業創新、完善監理規範及實現社會整體利益方面,產出具體且可衡量的成果。

