Woman-Run
In this trailer, the host Meredith Lowry sits down with Deanna Ray, the Chief Legal Officer for the Little Rock Reclamation Authority, to discuss a rapidly evolving corporate necessity: AI Policy. Triggered by the Arkansas State Legislature’s Act 848, public and government entities are now legally required to establish formalized AI policies if they utilize any form of artificial intelligence. Deanna breaks down the legalities, risks, and essential guardrails organizations must implement to protect themselves, their employees, and their clients in the age of automation. Deanna explains that the state legislature passed Act 848 because "water wastewater treatment actually is very technology-based," requiring a proactive stance on regulation. However, she clarifies that AI is defined broadly under the law, meaning it isn't just about generative platforms. It applies to "any artificial intelligence," which can include text autocorrect, automated note-takers, and spelling or grammar checkers. A critical element of any policy is the absolute requirement for human oversight. Deanna emphasizes that organizations "should not allow your AI technology... to make decisions for your entity without some human oversight". While AI can help narrow down choices or provide data insights, the human component must remain the ultimate decision-maker. The hosts also look closely at data retention, privacy, and the massive risks associated with third-party platforms. They discuss the necessity of having a plan to "scrape back" data and manage offboarding when an employee leaves the company. This ensures that an individual hasn't "put anything out there that could possibly cause security breaches or privacy breaches". Furthermore, they uncover surprising liabilities hidden in standard software terms of use. When discussing popular tools like Microsoft Co-pilot, they reveal a startling clause in the terms of use: "It is for entertainment purposes only". Because of this, the tech companies carry absolutely no liability if an employee makes a mistake based on the platform's outputs. Another major risk involves the threat to confidentiality and attorney-client privilege when using automated AI note-takers. Deanna warns that "the last thing you want to do is destroy your attorney-client privilege because you or your client has used some sort of note-taker via AI". Recent court cases have already shown that if a defendant asks a generative AI program about their legal case before hiring counsel, the government can successfully subpoena those inputs and outputs because they are not deemed privileged or confidential. Finally, the policy must outline strict rules for acceptable use, specifically targeting free versus paid platforms. The hosts remind listeners that if an organization relies on free tools, "you're probably because you're the product". Once data is entered into a free tool, an organization loses "complete control of all of it". To combat these issues and the growing threat of "automation bias," organizations must strictly define who has access to AI and keep critical thinking at the forefront, remembering the age-old tech rule: "Garbage in, garbage out". "We do have to have the policy by law. In 2025, the state legislature passed Act 848, which requires any government or public entity to have an AI policy if they utilize any sort of artificial intelligence." — Deanna Ray "You should not allow your AI technology, whatever version that you use, to make decisions for your entity without some human oversight... Ultimately, you as the human component need to be making decisions." — Deanna Ray "The last thing you want to do is destroy your attorney-client privilege because you or your client has used some sort of note-taker via AI." — Deanna Ray
24 episodios
Comentarios
0Sé la primera persona en comentar
¡Regístrate ahora y únete a la comunidad de Woman-Run!