Tuesday, February 27, 2024

AI use in California Education.

Karen M. Rezendes, Esq.
Managing Partner, Lozano Smith.
Lozano Smith, an education law firm, did a podcast today, regarding the use of AI in education in California, and I put some notes together when listening to it. The discussion concerns the impact and integration of generative AI technologies like ChatGPT, within public agencies and school districts. It was hosted by Karen Resendez, the managing partner at Lozano Smith, alongside Nick Clair, Rob Lomeli, and Karina Demirchian from the firm's artificial intelligence working group.

Generative AI, including systems capable of creating text, images, and videos, is spotlighted for its potential to mimic human reasoning and creativity. The conversation highlights the accessibility of AI tools, available in both free and paid versions, and stresses the crucial role of human oversight in ensuring the reliability and accuracy of AI-generated content. In practical applications, AI is being utilized in drafting ordinances and providing language translations for public meetings, as well as aiding educators and parents in developing Individualized Education Program (IEP) goals for students with disabilities. These examples underscore AI's utility in enhancing efficiency and inclusivity in educational and public services. However, the podcast also addresses significant concerns such as privacy, the risk of bias, and the phenomenon of AI "hallucinations," where the technology might produce inaccurate or fabricated information.

Legal considerations form a core part of the discussion, including the mention of a Biden Administration executive order aimed at regulating AI models and directing federal agencies, including the Department of Education, to develop regulations covering employment, bias, data privacy, and educational AI use within a year. Furthermore, about a dozen bills related to AI are being explored in the California legislature, indicating a proactive stance toward establishing guidelines on AI usage, its applications, and the types of AI systems that can be employed. The podcast reviews the potential legal risks and liabilities associated with AI use, including copyright issues and general liability for inaccuracies or harm resulting from AI-generated content. The panelists underscore the importance of public agencies maintaining a human in the loop for accountability and decision-making processes informed by AI.

Recommendations for public agencies include the establishment of policies and parameters for AI use, the importance of evaluating AI systems and use cases, and the necessity of hiring or consulting with experts to develop and adapt policies as AI technology evolves. Training for employees, staying abreast of developments in AI, and revising existing policies to address AI's impact are emphasized as critical steps for agencies. The concerns about privacy and bias are highlighted, advising against inputting personally identifiable information into AI systems, particularly free versions, and acknowledging the potential biases in AI responses due to the data on which they are trained.

The podcast is available at this link: Lozano Smith Podcast.