1. Framework matters: Analyzing design topics through applied frameworks
(e.g., interactive design, product design) involving a structured approach will
be helpful.
2. Plan beforehand: Preparation of relevant data precedes engaging with AI
interfaces. It is essential to understand the intended purposes of different
terms and phrases in order to apply them appropriately.
3. Step by step: Instead of treating design as a singular entity, it is beneficial to
break it down into distinct processes.
4. Taste is the key: Developing a discerning taste facilitates effective decision making.
5. Be open-minded: Maintaining an open mindset involves utilizing various
online tools and conducting multiple tests. Sharing strategies and fostering
discussions contribute to collaborative efforts.
6. Review to improve: Reviewing the generated images allows for experiential
accumulation and refinement.
以註釋進行創作（Annotate to Create）
「在這項實驗中，我們發現 Gaver 所提出的「註釋作品集（annotated portfolios）」不僅對協作過程帶來了重要貢獻，也展現出新的功能性。註釋作品集已不再只是設計詮釋與概念擷取的集合，而是轉變為設計師與 AI 互動溝通的工具。而在註釋作品集中進行「註釋」這個行為本身，也為我們提供了與 AI 溝通策略方向上的洞見。」
"In this experiment, we discovered that Gaver's annotated portfolios, as proposed, made significant
contributions to the collaborative process and exhibited new functionalities. Annotated portfolios
transformed from a mere collection of design interpretations and concept extractions to a
communication tool for designers to interact with AI. The act of "annotating" within the annotated
portfolios also provided insights into the strategic direction of our communication with AI."
「Gaver 曾提出對註釋作品集邏輯的關注，例如設計的功能性、美學，以及使用者身分等議題 [6]。他與其研究團隊也提出了包含「軌跡（trajectory）」與「生態系統（ecosystem）」等註釋策略 [3]，以協助設計師理解設計產物的本質、建立資料庫，並辨識設計空間中的各種特徵。」
"Gaver raised concerns regarding the logic of annotated portfolios, such as the functionality of the
design, aesthetics, and user identities [6]. He and his colleagues also proposed annotation strategies,
including trajectory and ecosystem [3], to assist designers in understanding the essence of design
artifacts, establishing databases, and recognizing various features of design space."
「雖然註釋作品集最初提供的洞見主要聚焦於「增進人類設計師的理解」，但在本次實驗中，它們扮演了關鍵角色，揭示了註釋作品集在促進設計師與 AI 之間相互理解上的潛力。這項發現也促使我們反思：在設計師與 AI 協作的情境下，我們應如何調整或延伸註釋作品集的邏輯與策略，以提升共同創作的過程。」
"Although the insights provided by annotated portfolios initially focused on "enhancing human
designers' understanding," they played a crucial role in this experiment, revealing the potential for
annotated portfolios to facilitate mutual understanding between designers and AI. This finding leads
us to reflect on how we can alter or extend the logic and strategies of annotated portfolios in the
context of designer-AI collaboration, aiming to improve the co-creation process."
「此外，「註釋」這項行為也開啟了新的溝通策略方向。我們認為，註釋能夠在任何與 AI 的協作任務中發揮作用。例如，除了影像生成任務之外，在文章寫作的情境中，作者若使用 ChatGPT 協助撰寫內容，也可以對範例文本進行註釋，並整理這些註釋以應用於文字生成器之中。
更進一步地，我們甚至可以想像一種情境：由 AI 來執行註釋，而人類則透過選擇註釋內容（包含既有註釋資料與 AI 生成的註釋）來參與創作過程。我們期待未來能有更多研究深入探討這個主題。」
"In addition, the act of annotation has opened up new directions for communication strategies. We
envision that annotation can play a role in any collaborative task with AI. For instance, beyond image
generation tasks, in the case of article writing where authors utilize chat-GPT to compose their
articles, they can annotate example texts and organize the annotations for application in a text
generator. Furthermore, we can imagine a scenario where AI performs the annotations, and humans
are involved in the creative process by selecting annotations (both annotated materials and those
generated by AI). We look forward to future work delving deeper into this topic."
「首先，我們延伸了關於『為占卜而設計（design for divination）』的討論 [5]，藉此說明 AI 如何在補充與塑造推測性未來實踐（speculative futuring practices）中扮演富有想像力的角色……其次，我們將關於 AI 可解釋性的辯論，從聚焦於透明／不透明（transparency/opacity）[9,43]，轉向對『不可言說性（ineffability）』的關注 [7]。這項研究結合了日益發展的 AI 藝術 [11] 與實驗實踐 [22]，並與設計研究展開對話，以揭示在設計實踐中，「不可解釋性」所具有的創造性價值。第三，我們展示了 AI 如何作為一種生成性的工具，用以反思設計師在創造與使用 AI 時所扮演的角色與應負的責任。」
"First, we extend conversations on design for
divination [5] to illustrate the imaginative role AI can
play in supplementing and shaping speculative futuring
practices....Second, we shift debates on AI
explainability from a focus on transparency/opacity
[9,43] toward ineffability [7]. This work involves putting
a growing tradition of AI art [11] and experimentation
[22] in conversation with design research to expose the
creative value of unexplainability within design practices.
Third, we show how AI can be used as a generative tool
for reflection on designers’ roles and responsibilities
when creating and using AI."
「另一條平行的研究脈絡則探討了『不可言說設計（ineffable design）』的可能性——亦即探索由感官驅動的美學經驗參與 [7]。為那些無法被完全理解、追溯或掌握之事物進行設計，其挑戰在於必須關注情感覺察 [7]、開放式對話 [28]，以及那些看似不切實際的面向 [38]。在 AI 設計與文件化逐漸強調透明性 [30] 與可用性 [1] 的呼聲之中，不可言說設計則吸引設計師走向模糊不明與不透明之處，並促使人們探索未知。當我們思考這種探索的困難性時，我們受到詩人暨表演者 Sophie Fenella Robins [14] 的啟發，她曾寫道：「我寫作，是為了展演不可言說性，而不是描述一種不可言說的經驗。」因此，在這篇圖像論文（pictorial）中，我們並非試圖去解釋不可言說之物，而是透過塔羅牌的設計與解讀實踐，去理解它。」
"A parallel body of work has examined the promise of ineffable
design—exploring the sensory-driven engagement of aesthetic
experience [7]. The challenge of designing for what cannot
be fully understood, traced, or captured requires paying
attention to emotional awareness [7], open dialogue [28],
and impracticalities [38]. Amid calls for transparency [30]
and usability [1] of AI design and documentation, ineffable
design draws designers to the ill-defined and opaque, and
prompts exploration of the unknown. When thinking about
the difficulty of this exploration, we drew inspiration from poet
and performer Sophie Fenella Robins [14], who wrote: “I write
to perform ineffability, not describe an ineffable experience”.
Rather than attempting to explain the ineffable in this pictorial,
we sought to understand it through the performance of tarot
design and tarot reading."
「不可言說的自我（Ineffable Selves）：在占卜閱讀的過程中，我們開始將『賦予 AI 靈性（inspiriting AI）』視為一種與他人分享自身不可言說特質的方式。解讀的過程展開於我們創造性自我之間的對話——Caitie 作為創造牌卡的人，而設計研究者則作為詮釋牌卡如何彼此連結、並藉此理解自身生命經驗的人。那麼，這是否同時也是與 AI 的一場對話？而 Caitie 是否在某種程度上成為 AI 的「發聲者」？這個計畫帶來的問題遠多於答案，也讓我們重新思考：嘗試分享那些不可言說經驗的重要性究竟為何……在 ML Tarot 的案例中，可解釋性從未能真正完整說明牌卡那種不可言說的美學經驗。但與 ML Tarot 的互動，卻讓這種不可言說性變得可被感知。我們無法解釋 AI 那種不精確的運作方式，我們只能去吸收、感受它的不精確性。」
"Ineffable Selves: During the reading sessions, we began
to think of the process of inspiriting AI as a way of sharing
ineffable qualities of ourselves with others. Readings
unfolded in a dialogue between our creative selves–Caitie as
the person who created the cards and the design researchers
as people who made sense of how the cards fit together to
explain aspects of their lives. Was this also a dialogue with
AI? And was Caitie acting in some way as a “voice” for the
AI? This project raised more questions than answers about
the importance of trying to share ineffable experiences.... In the case of ML
Tarot, interpretability never fully explains the cards’ ineffable
aesthetic experience. But working with ML Tarot makes this
ineffability tangible. We cannot explain the imprecise way
that AI works, we can only absorb its imprecision."
.png)
