Human Centered AI Designing for Impact and Responsibility #254031

Upon completion of this course, participants will be able to:

• Articulate the Principles of HCAI: Define Human-Centered AI and differentiate it from technology-driven approaches, understanding its core tenets of fairness, accountability, transparency, and human wellbeing.

• Apply Human-Centered Design Methodologies: Integrate HCD practices (e.g., stakeholder mapping, participatory design, value-sensitive design) into the AI development lifecycle.

• Identify and Mitigate Bias: Implement technical and procedural strategies to audit for, detect, and mitigate bias and discrimination in AI systems at every stage.

• Design for Transparency and Explainability (XAI): Create AI systems that provide meaningful explanations for their outputs, fostering user trust and enabling effective human oversight.

• Evaluate AI Impact and Risk: Conduct pre- and post-deployment impact assessments to proactively identify potential social, ethical, and economic harms.

• Develop Governance Frameworks: Establish organizational processes, guidelines, and review boards for the responsible development and deployment of AI.

This course is designed for a multidisciplinary audience involved in the creation, governance, and application of AI systems:

• AI Product Managers & Designers (UX/UI)

• Data Scientists & Machine Learning Engineers

• Software Developers & Engineers building AI-powered applications

• Ethics, Risk, & Compliance Officers

• Policy Makers & Regulatory Affairs Specialists

• Corporate Strategists & Innovation Leaders

• Social Scientists & Researchers studying technology’s impact

• Legal Counsel specializing in technology and privacy

• Pre-assessment
• Live group instruction
• Use of real-world examples, case studies and exercises
• Interactive participation and discussion
• Power point presentation, LCD and flip chart
• Group activities and tests
• Each participant receives a binder containing a copy of the presentation
• slides and handouts
• Post-assessment

Day 1: Foundations of Human-Centered AI

• AM: Why HCAI? The Imperative for a New Approach

  • The limitations of tech-centric AI: Case studies of AI failures and public backlash.

  • Defining Human-Centered AI: Principles and parallels to Human-Centered Design.

  • The multi-disciplinary team: Why ethicists, social scientists, and domain experts are essential.

• PM: The HCAI Framework & The AI Lifecycle

  • Mapping the extended AI lifecycle from ideation to decommissioning through a human-centric lens.

  • Introduction to key frameworks: Value-Sensitive Design (VSD) and Participatory Design.

  • Workshop: Identifying stakeholders and human values for a sample AI project.

Day 2: Designing for Fairness and Mitigating Bias

• AM: Understanding Algorithmic Bias

  • Types of bias: historical, representation, measurement, and aggregation bias.

  • The societal context of data: How bias gets embedded in datasets and models.

  • Case Study: Analyzing a real-world example of discriminatory AI.

• PM: The Practice of De-biasing

  • Technical mitigations: Pre-processing, in-processing, and post-processing techniques.

  • The role of diverse teams and external audits.

  • Hands-on Exercise: Using an open-source toolkit (e.g., IBM AIF360 or Google’s What-If Tool) to analyze a dataset for bias.

Day 3: Building Trust through Transparency and Explainability

• AM: The “Black Box” Problem

  • The spectrum of explainability: from global model understanding to local explanations.

  • Explainable AI (XAI) techniques: LIME, SHAP, and counterfactual explanations.

  • Designing explanations for different audiences (users, developers, regulators).

• PM: Designing for User Agency and Control

  • Principles of contestability: Allowing users to challenge and correct AI decisions.

  • Incorporating user feedback loops for continuous learning and improvement.

  • Workshop: Storyboarding an interface that explains an AI decision and offers a recourse path.

Day 4: Implementing Governance and Measuring Impact

• AM: AI Ethics Governance in Practice

  • Creating an AI Ethics Board or Review Committee.

  • Developing organizational guidelines and checklists for responsible AI.

  • The role of AI Impact Assessments (e.g., based on the EU AI Act’s requirements).

• PM: Measuring What Matters: Beyond Accuracy

  • Defining success metrics for HCAI: fairness, wellbeing, trust, and long-term societal impact.

  • Methods for ongoing monitoring and auditing of deployed models.

  • Group Exercise: Drafting an impact assessment for a proposed AI system.

Day 5: Synthesis and Action

• AM: The Business Case for Responsible AI

  • How HCAI manages risk, builds brand trust, and creates long-term value.

  • Communicating HCAI principles to executives, investors, and customers.

  • Overcoming organizational barriers to implementing HCAI practices.

• PM: Capstone Project: The HCAI Blueprint

  • Final Exercise: In teams, participants will develop a comprehensive HCAI blueprint for a new AI application. This will include:

    • Stakeholder and value analysis

    • Bias mitigation strategy

    • Explainability and UI design plan

    • Governance and impact assessment plan

  • Presentations and critique session.

  • Course Wrap-Up: Resources for continued learning and building a community of practice.