全球都在搶最強AI，台灣為什麼反而更需要自己的？| 思想實驗室 Video Podcast ep124：ft.台智雲 吳漢章總經理

A recent discussion highlights the complexities and challenges surrounding artificial intelligence (AI) in Taiwan, particularly in light of a controversial incident involving a service robot at a five-star hotel that referred to Taiwan as a province of China. This incident raises questions about the biases inherent in AI systems and the importance of understanding the sources of their information. Experts emphasize that the responsibility for AI outputs lies not only with the models themselves but also with the operators who select and configure them.

Peter Wu, General Manager of Taichung Cloud, elaborates on the need for a robust framework to ensure that AI systems are trained on reliable data and that their outputs are validated by domain experts. He notes that as AI becomes more integrated into sectors like healthcare and finance, the stakes are higher, necessitating a careful approach to model selection and deployment. The conversation also touches on the regulatory landscape, with Taiwan recently passing its own AI framework to manage risks associated with AI technologies.

The discussion further explores the growing demand for "sovereign AI," which emphasizes local control over data and technology. Wu expresses optimism about Taiwan's potential to develop its own AI capabilities, driven by its strong hardware industry and the need for localized solutions. He suggests that as AI technology evolves, the focus should be on creating systems that are not only effective but also trustworthy and accountable, ensuring that users can rely on AI without compromising their data security.

Overall, the dialogue reflects a broader trend in Taiwan towards embracing AI while navigating the associated risks and ethical considerations, highlighting the importance of local expertise and regulatory support in fostering a responsible AI ecosystem.

投資理財要找合法金融機構！ 保證獲利不要信、陌生群組不要加、退休理財要穩健、契約文件要看清，守護財務安全你也做得到。 以上資訊，由財團法人金融消費評議中心提供。 -- 選舉紅包一直加，公托名額落誰家？ 台灣前進盤點：明年地方加碼福利，長輩623億、青壯年14億，每人平均差逾百倍！支持照顧長輩，也要看見青壯年。 上台灣前進官網看紅包排行，追問財源與世代公平：誰買單？ https://sofm.pse.is/9sabug －－－－以上為 SoundOn 動態廣告－－－－ 當 AI 越來越厲害，你也越來越依賴，但當模型漲價、規則改變、服務斷線時，你怎麼辦？

AI 會講繁體中文，不代表它真的懂台灣； 資料存在台灣，也不代表 AI 不會出境； 取得 GPU，更不代表擁有「自己的 AI」。

當 AI 開始進入醫療、金融、企業營運，關鍵不只是「哪個模型最強？」，而是你到底掌握多少控制權與選擇權？ 本集《思想實驗室》邀請台智雲（Taiwan AI Cloud）總經理吳漢章 Peter，從模型、算力一路談到 AI Agent，拆解「主權 AI」為什麼正在從國家級議題，成為每家企業都必須關注的重要問題。

也一起聊聊：台灣已經替全世界製造最重要的 AI 硬體，下一步，有沒有機會將這項優勢發展為掌握在自己手上的 AI 能力？

👉 了解台智雲 Model Foundry: https://bit.ly/483dQKm 👉 台智雲主權 AI 模型進 AIEC 榜單，350 億參數以下名列第 2、第 3 名: https://bit.ly/4yoMB7P

// 製作團隊 // 主持：泛科知識 #鄭國威知識長 企劃：廖儀瑄 剪輯：GINA

🎙 思想實驗室 Video Podcast 每週帶你探索最新科技、科學議題與能源趨勢！訂閱我們，記得按讚、訂閱，並開啟小鈴鐺，不錯過每一集！ #主權AI #AI算力 #人工智慧 #國家韌性 #思想實驗室#人工智慧 #AI安全 #ChatGPT #台灣AI -- Hosting provided by SoundOn