Jascha Goltermann: The Impact of AI on UX Design - Hatch Conference 2023

Summary

Yasha Golderman's talk at a conference delves into the transformative impact of AI on UX design, using Booking.com's AI Trip Planner as a case study. He discusses the evolution from traditional UX design to incorporating generative AI, emphasizing the necessity for designers to understand AI's capabilities and limitations. Golderman highlights the importance of prompt engineering, setting AI guardrails, and maintaining user trust and control, while advocating for designers to stay empathetic and adapt to the changing landscape of AI-enhanced design.

Takeaways

• 🚀 The speaker emphasizes the growing interest in managing people and the nonlinear ways of managing a career in the tech industry.
• 🌟 Yasha Goldman's talk focuses on the impact of AI on UX design, specifically how AI is integrated into products rather than design tools.
• 🔍 AI's presence is exemplified by everyday applications like image recognition and search functionalities in smartphones, showcasing its accessibility and integration into user experiences.
• 🏢 Yasha's background in UX for over a decade and his role at Booking.com, where he manages designers and leads projects like the AI Trip Planner, sets the stage for his expertise on the subject.
• 🌐 Booking.com's scale, with 700 million monthly visitors and a strong data-informed approach, underscores the company's user-centric mindset and the potential for AI applications.
• 🤖 The distinction between AI, machine learning, deep learning, and generative AI (Geni) is clarified, highlighting the latter's role in creating diverse data types.
• 💬 The integration of Geni into everyday tools and platforms like Notion and Grammarly indicates a shift in user expectations and the need for designers to adapt.
• 🛍️ The AI Trip Planner at Booking.com is presented as a case study, illustrating how AI can enhance services like virtual travel agents with more natural and context-aware interactions.
• 🛑 The importance of setting guardrails for AI to ensure appropriate output and maintain brand values is stressed, especially in avoiding sensitive topics and malicious intent.
• 🔄 The need for designers to shift focus from UI to AI, considering the quality of AI output as a defining factor in user experience, is highlighted.
• 🔄 The talk concludes with a call for designers to understand AI, engage in prompt engineering, and embrace the concept of hyper-personalization to stay relevant in a rapidly evolving field.

Q & A

What was the main topic of Yasha Golderman's talk at the conference?

Yasha Golderman's talk focused on the impact of AI on UX design, specifically discussing how AI can be integrated into products to improve user experiences, using the AI trip planner at Booking.com as a case study.

What is the difference between AI, machine learning, and deep learning as mentioned in the script?

AI refers to the broader field of study of artificial intelligence. Machine learning is a subset of AI where computers are trained on large datasets to find patterns and make decisions. Deep learning takes this further by creating artificial neural networks with multiple layers to mimic the human brain, allowing the computer to learn more complex patterns.

What is the role of generative AI (gen AI) in the context of the talk?

Generative AI, or gen AI, is a type of AI that can create new data such as text, images, or video. In the talk, gen AI is highlighted for its potential to enhance user experiences in products, with examples of how it's already integrated into everyday tools and the Booking.com AI trip planner.

How has Booking.com utilized AI in its services prior to the development of the AI trip planner?

Booking.com had previously developed the Booking Assistant, which used machine learning models to help customers with trip-related questions. However, it was limited to specific topics and did not understand broader conversation contexts or apply real-world knowledge until the integration of gen AI.

What is the AI trip planner, and how does it differ from the previous Booking Assistant?

The AI trip planner is an enhancement of the Booking Assistant, integrated with open AI's GPT model, allowing it to discuss any travel-related topic in a more natural way, understand broader conversation contexts, and apply real-world knowledge to provide more personalized and accurate recommendations.

What is the significance of the number of visitors, listings, and reviews that Booking.com has in relation to AI?

The large volume of data from Booking.com's 700 million monthly visitors, nearly 30 million global listings, and over 250 million guest reviews provides a rich dataset for AI to learn from and offer personalized recommendations, enhancing the AI's ability to understand and cater to user needs.

Why is prompt engineering important for designers working with AI?

Prompt engineering is crucial as it involves instructing the AI on how to respond appropriately to user inputs. It helps in setting the AI's personality, defining guard rails, and ensuring the AI's responses align with the brand's values and user expectations.

How should designers approach the design process when working with gen AI?

Designers should shift their focus from UI to AI, mapping situations rather than steps, as gen AI interactions are nonlinear. They need to consider user flows that can adapt to various AI responses and ensure the AI's output is of high quality and meets user expectations.

What are some of the ethical considerations when integrating AI into a product?

Designers must consider setting guard rails to prevent inappropriate outputs, respecting sensitive subjects, and accounting for potential malicious intent from users trying to push the AI beyond its intended limits.

How can designers ensure users feel in control when interacting with AI-enhanced products?

Designers can ensure users feel in control by not using their data to train the AI models, allowing users to delete their conversations, and providing clear examples of how to interact with the AI, as well as managing expectations through disclaimers and welcome messages.

What are some resources recommended in the script for designers to learn more about AI and prompt engineering?

The script recommends Ben B's newsletter and the Invisible Machines podcast for staying updated on AI, the Prompt Engineering Guide for learning about instructing AI, and the Delo report 'Connecting with Meaning' for understanding hyper-personalization.