Sketching NLP

Yang, Qian; Cranshaw, Justin; Amershi, Saleema; Iqbal, Shamsi T.; Teevan, Jaime

doi:10.1145/3290605.3300415

Cited by 59 publications

(24 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This approach enables HCI professionals to rapidly explore many design possibilities and probe user behaviors. However, this approach fails to address any of the UX issues that will come from AI inference errors because there is no way to simulate these errors [52]. The second approach is to create a functioning AI system, and deploy it among real users for a period of time [53].…”

Section: Ux Design Challenges Of Aimentioning

confidence: 99%

“…• Challenges in collaborating with AI engineers (throughout the design process: For many UX design teams, AI technical experts can be a scarce resource [19,53]. Some designers also found it challenging to effectively collaborate with AI engineers, because they lacked a shared workflow, boundary objects, or a common language for scaffolding the collaboration [19,28,52].…”

Section: Ux Design Challenges Of Aimentioning

confidence: 99%

“…First, we draw on our many years of experience in designing a wide range of AI systems, from simple adaptive user interfaces [58,12], to large-scale crowd-sourced transportation systems [59]; from clinical decision supports [57,55] to natural language productivity tools [52,60,16]. These experiences enabled us to give a firsthand account of the design challenges of AI, as well as a felt understanding of the solution that naturally emerged from the process.…”

Section: Designing the Ux Of Ai Applicationsmentioning

confidence: 99%

“…For instance, designers struggled to understand what AI can and cannot do even when they understood how AI works [13]; This is because the capabilities of an AI system can be inherently uncertain and constantly evolving. Designers struggled to rapidly prototype human-AI interaction [52] because the interactions of two mutually adaptive agents resist easy abstraction or simulation. Designers struggled to follow a typical user-centered design workflow when designing human-AI interactions [19,51].…”

Section: Two Complexity Sources Taken Togethermentioning

confidence: 99%

See 3 more Smart Citations

Re-examining Whether, Why, and How Human-AI Interaction Is Uniquely Difficult to Design

Yang

Steinfeld

Rosé

et al. 2020

Proceedings of the 2020 CHI Conference on Human Factors in Computing Systems

Self Cite

336

182

View full text Add to dashboard Cite

Artificial Intelligence (AI) plays an increasingly important role in improving HCI and user experience. Yet many challenges persist in designing and innovating valuable human-AI interactions. For example, AI systems can make unpredictable errors, and these errors damage UX and even lead to undesired societal impact. However, HCI routinely grapples with complex technologies and mitigates their unintended consequences. What makes AI different? What makes human-AI interaction appear particularly difficult to design? This paper investigates these questions. We synthesize prior research, our own design and research experience, and our observations when teaching human-AI interaction. We identify two sources of AI's distinctive design challenges: 1) uncertainty surrounding AI's capabilities, 2) AI's output complexity, spanning from simple to adaptive complex. We identify four levels of AI systems. On each level, designers encounter a different subset of the design challenges. We demonstrate how these findings reveal new insights for designers, researchers, and design tool makers in productively addressing the challenges of human-AI interaction going forward.

show abstract

Section: Ux Design Challenges Of Aimentioning

confidence: 99%

Section: Ux Design Challenges Of Aimentioning

confidence: 99%

Section: Designing the Ux Of Ai Applicationsmentioning

confidence: 99%

Section: Two Complexity Sources Taken Togethermentioning

confidence: 99%

See 2 more Smart Citations

Re-examining Whether, Why, and How Human-AI Interaction Is Uniquely Difficult to Design

Yang

Steinfeld

Rosé

et al. 2020

Proceedings of the 2020 CHI Conference on Human Factors in Computing Systems

Self Cite

336

182

View full text Add to dashboard Cite

show abstract

“…As new technologies such as natural language processing become a core design material within interaction design, pressures to avoid premature costly data collection and technical development are likely to exacerbate interest in WOz, as can be seen from a string of recent publications in HCI-based research (e.g. [3,28,33,45,46,49]). This work often describes the use of WOz somewhat cursorily, glossing over its application.…”

Section: Wizard Of Oz In Hci and Cscwmentioning

confidence: 99%

Pulling Back the Curtain on the Wizards of Oz

Porcheron

Fischer

Reeves

2021

Proc. ACM Hum.-Comput. Interact.

View full text Add to dashboard Cite

The Wizard of Oz method is an increasingly common practice in HCI and CSCW studies as part of iterative design processes for interactive systems. Instead of designing a fully-fledged system, the 'technical work' of key system components is completed by human operators yet presented to study participants as if computed by a machine. However, little is known about how Wizard of Oz studies are interactionally and collaboratively achieved in situ by researchers and participants. By adopting an ethnomethodological perspective, we analyse our use of the method in studies with a voice-controlled vacuum robot and two researchers present. We present data that reveals how such studies are organised and presented to participants and unpack the coordinated orchestration work that unfolds 'behind the scenes' to complete the study. We examine how the researchers attend to participant requests and technical breakdowns, and discuss the performative, collaborative, and methodological nature of their work. We conclude by offering insights from our application of the approach to others in the HCI and CSCW communities for using the method.

show abstract