Key Takeaways 1. AI adoption should not be measured only by efficiency Legal teams often focus on time saved, headcount reduction, and cost containment. Marisa argues these are incomplete metrics. Organizations also need to assess AI’s effects on human well-being, discernment, decision-making, and cognitive capacity. 2. Humans are not fully “ready” for AI Marisa introduces the idea of human readiness — including cognitive, psychological, and social readiness. AI may be advancing faster than people’s ability to responsibly process and evaluate its outputs. 3. Cognitive overload leads to offloading judgment When users are overwhelmed by too much information, they unconsciously defer to the machine. In legal practice, that can mean diminished critical thinking, less discernment, and increased reliance on outputs that may be wrong but appear convincing. 4. AI bias isn’t only about race or gender Marisa highlights lesser-discussed cognitive biases such as automation bias, authority bias, and the privacy paradox. These biases shape how people interact with AI systems and can undermine judgment. 5. AI systems are designed for continuous engagement One of the episode’s most striking points: many AI interfaces are built to keep users engaged, much like social platforms or slot-machine mechanics. That raises serious questions about manipulation, dependency, and user autonomy. 6. Training is the answer — but not just tool training Organizations need more than prompt training. They need training in how AI affects thinking, relationships, judgment, and work habits. Human readiness requires practice, reflection, and institutional support. 7. Legal professionals have a unique role to play Lawyers are trained to parse complexity, evaluate facts, and question assumptions. Marisa suggests the legal profession is well-positioned to help establish stronger practices around AI accountability and safe use. 8. There is a serious client confidentiality risk in consumer AI tools Marisa raises a practical concern: clients and legal staff may unknowingly upload sensitive information into public AI tools. This creates risks around admissibility, confidentiality, and competence. 9. AI has significant environmental consequences The conversation also explores the strain AI places on the electrical grid, carbon emissions, and fresh water resources used by data centers — an underdiscussed but growing area of concern. 10. The goal is not fear — it’s discernment Marisa is not anti-AI. Her core message is that AI can be beneficial, but only if people learn how to use it without surrendering agency, judgment, or responsibility. Marisa Zalabak | LinkedIn Guest Notes Marisa Zalabak co-authored the IEEE 7010 standard which addresses recommended practices for assessing the impact of autonomous and intelligent systems on human well-being. She also chairs the IEEE AI Ethics Education Committee, works on standards related to emulated emotions in AI systems, leads the Planet Positive 2030 Initiative, and is the founder of Open Channel Culture. Her work focuses on the intersection of AI ethics, human development, education, cognition, and societal impact.
