Using Al Ethically to Review Cases Date: Wednesday, February 12 2025 Maximizing the benefits of Al - Familiarization and training • Continuous learning is key! - Understanding limits and capabilities - Choosing the right tools for specific tasks - Using Al for routine and repetitive tasks - Independently verify and review Al output - Ethical considerations and client communications - Supervisory responsibilities - Al as a supplement - not a replacement! Confidential Client Information and Al - - Understand the Al Tool's data handling • Thoroughly review the Terms of Use and other relevant documentation • Look for a "Trust Center" on their website Ensure the provider does not share inputted information with third parties or use it for its own training or product improvement - Anonymize client information - Consult with your IT team and/or cybersecurity experts - Avoid public/free platforms for processing sensitive information Open source VS Open weight - Open source: Provides complete access to all components needed for reproducibility • - Includes access to training data, model architecture, and ability to retrain Open weight: Allows users to freely download and deploy Al models without charge • Does not include access to training data or ability to fully retrain/modify the model - Closed/proprietary: Models with no access to weights or training data More Helpful Terms - - RAG (Retrieval-Augmented Generation): • Gives Al access to a knowledge base of documents/data it can search through • Improves output quality and reduces hallucination Agentic Al • Al tools that can actively take actions on behalf of the user, like using tools • More of an actual "assistant" than a "chatbot" Al Experts To Follow - Ethan Mollick: • Professor at the Wharton School of the University of Pennsylvania, specializing in innovation and entrepreneurship. • Recognized for his research on the impact of Al on work and education, as well as for his practical insights into how Al can be effectively integrated into various domains. • Author of Co-Intelligence: Living and Working with Al, outlines principles for collaborating with Al technologies, making him a leading voice in understanding the future of human-Al interaction • - Follow on Substack, LinkedIn, BlueSky, etc. Allie Miller: • previously served as the Global Head of Machine Learning Business Development for Startups and Venture Capital at Amazon Web Services (AWS) • advocate for diversity in technology, she co-founded initiatives like "Girls of the Future" and has spoken at over 100 global events on topics related to Al implementation and ethics • Follow on LinkedIn, Instagram, etc.