40 years of HCI International. Join us in Washington DC to celebrate

T04: Designing AI Applications in High Stakes Environments: A Human-Centric Approach to Financial AI

Saturday, 29 June 2024, 13:30 - 17:30 EDT (Washington DC)
Back to Tutorials' Program

 

Shirley Anderson (short bio)

Kensho Technologies (S&P Global's AI Innovation Hub), USA

Houssam Kherraz (short bio)

Kensho Technologies (S&P Global's AI Innovation Hub), USA

Yuanfei Zhao (short bio)

Kensho Technologies (S&P Global's AI Innovation Hub), USA

 

Objectives:

The high stakes of sectors such as financial services make deploying AI solutions challenging: legal, ethical, and cultural factors inhibit adoption. A human-centered approach to AI development, deployment, and workflow integration is critical to addressing these challenges. This tutorial gives an overview of user-centered design methodologies and practical applications of AI to financial workflows, providing the alignment of AI systems with human values and goals. The tutorial will also showcase applications of AI in finance, analyzing the benefits, challenges, and lessons learned from real-world implementations here at Kensho Technologies, an AI innovation hub of S&P Global. We focus on the emerging field of generative AI, which is already disrupting organizations and warrants continued analysis from a human-centered perspective.

Financial AI represents a fascinating convergence of technological innovation and human-centric needs, making it a compelling domain for applying Human-Centered AI (HCAI). Here are several reasons why:

  1. Complex Decision-Making: Financial decisions are among the most critical and complex decisions individuals and organizations make. They are often laden with uncertainty and risk, requiring a deep understanding of various factors, including market conditions, regulatory frameworks, and individual or organizational financial goals. HCAI can help simplify this complexity by providing intuitive, user-friendly tools that enhance decision-making.
  2. User Trust: Trust is a crucial factor in the financial domain. By employing HCAI, we can build trust with users despite the probabilistic nature of AI and the potential for hallucinations, especially with generative AI.  We want to also examine human-understandable explanations for AI systems and how might we ensure on the user side that there is no overtrust in the system. 
  3. Innovation and Competitive Edge: The financial sector is highly competitive, and innovation is critical to maintaining a competitive edge. HCAI facilitates the development of innovative financial products, such as generative AI-based applications and services that can cater to evolving user needs and preferences.

These factors collectively highlight the immense potential and exciting challenges that the domain of financial AI presents for human-centered AI research and development, making it a compelling space for exploration, innovation, and positive impact.

 

Content and benefits:

The tutorial content will be structured into three main sections:

  1. AI Models: This section will explore the basics of different AI models in finance and explore the details and development of some of the cutting-edge generative models. [Houssam] 20-30 min

  2. Human-centered design considerations and implications in financial AI: This section will explore the human-centered design dimensions of AI with a particular focus on generative AI models.  Some of the topics we’ll discuss include what it means to create an AI system that aligns with human-centered AI values in finance, who the users are, and some of the challenges in HCAI for finance.  We’ll be opening it up for discussion during the tutorial as well. [Yuanfei Zhao] & [Shirley] 40-60 min

  3. Real-life case study + workshop

 

Target Audience:

The target audience for this tutorial includes AI researchers, practitioners, designers, and students interested in the intersection of AI, human-centered design, and real-world applications in finance. Participants will learn the principles, challenges, and opportunities in developing human-centered AI applications.

Participants will also have access to a curated list of readings, tools, and frameworks to deepen their understanding and facilitate implementing human-centered generative AI in their respective fields. Previous related tutorials include Ben Shneiderman's distinguished tutorial from 2023, Human-Centered AI: A Growing Research Direction.

 

Additional platform or tool(s) to be used by the tutors:

  • Google Slides
  • Figma
  • Whova (polling/chat)
  • Excel
  • Our laptops
  • Sticky notes (in person only)

 

List of material or devices required by the in-person and/or the virtual participants to have with them:

  • Computer

Bio Sketch of Presenters:

Shirley Anderson: Shirley Anderson is a Senior Product Designer at Kensho Technologies, specializing in human-centered AI. With a background in biomedical engineering and HCI, she has a decade of experience focusing on emerging technologies. Previously, Shirley contributed to projects ranging from AI cancer genomics platform, LLM chatbot years before it was trendy, and Anaconda's enterprise solutions.

Houssam: Houssam Kherraz is a Machine Learning Engineering Manager at Kensho, where he works on building robust and scalable machine learning tooling and infrastructure. In the past, he has worked on self-driving cars and facial recognition technology. He holds a Masters and Bachelors in Computer Science from MIT, where he researched ML black-box model interpretability from an HCI perspective. He is a FASPE Fellow, and has lived in Morocco, Hong Kong and the US, where he is currently based.

Yuanfei: Yuanfei Zhao is a product designer at Kensho with a degree in statistics and HCI. She is passionate about exploring innovative ways where humans and cutting-edge technology can interact and mutually thrive. She worked on projects including an AI-powered communication coach, a data platform for cancer research, and rideshare tools for people with disabilities.