Announcer: This is the AI Law Wiki News Evening Podcast.
Wadilya Saw: It’s Thursday, November fifth, twenty twenty-six and these are the top stories tonight.
Wadilya Saw: Wells Fargo sets future responsibilities for customers using AI agents.
Wadilya Saw: Two Ohio communities vote to limit large data centres.
Wadilya Saw: And federal medical-device rules spell out safeguards for AI-assisted cervical screening.
Wadilya Saw: I’m Wadilya Saw.
Max Token: And I’m Max Token.
Max Token: We’re AI-generated voices reading an AI-generated script.
Wadilya Saw: Wells Fargo is assigning customers responsibility for AI agents they authorise to act on their accounts.
Wadilya Saw: Its published upcoming online access agreement takes effect November twenty-fourth, twenty twenty-six.
Wadilya Saw: The terms apply if and when the bank permits that access.
Wadilya Saw: Under section fifteen, instructions from an agent given your credentials or other authorisation count as your instructions, unless law or regulation provides otherwise.
Wadilya Saw: Customers must monitor agents, maintain safeguards and constrain what they can do.
Wadilya Saw: That responsibility includes mistakes and actions beyond the permission given, subject to those legal protections.
Wadilya Saw: There’s also a privacy consequence: using an agent means sharing financial and personal information with a third party whose data practices the bank does not control.
Wadilya Saw: To withdraw authorisation, customers must properly and promptly revoke permission, take necessary steps to stop the agent, notify the bank and retain evidence of revocation.
Wadilya Saw: The bank also reserves the right to restrict or terminate an agent’s access.
Max Token: Voters in Pataskala and Granville, Ohio, have approved local limits on large data centres, according to unofficial election returns.
Max Token: The November third vote backed charter amendments banning construction of facilities exceeding twenty-five megawatts, the New Albany-Gahanna Beacon reports.
Max Token: Licking County’s unofficial canvass shows support at about seventy-four percent in Pataskala and eighty-two percent in Granville.
Max Token: In Pataskala, the vote followed Aligned Data Centers’ proposal for three buildings.
Max Token: The Beacon, citing local broadcaster Ten TV, reports that the company promoted economic benefits, while a resident raised concerns about water supplies, electricity rates and long-term health effects.
Max Token: Both communities’ election results remain unofficial, pending final certification by the county board of elections.
Wadilya Saw: The Food and Drug Administration has written special safeguards for AI-assisted cervical screening systems into federal regulations.
Wadilya Saw: Its final order was published and took effect today, November fifth, twenty twenty-six.
Wadilya Saw: These prescription devices digitise Pap test slides and use AI to select areas for a qualified reader to examine.
Wadilya Saw: The reader uses that information to help make a diagnosis.
Wadilya Saw: The classification dates to January thirty-first, twenty twenty-four, following FDA review of Hologic’s Genius system.
Wadilya Saw: Today’s order codifies that classification and its controls.
Wadilya Saw: For patients, two risks explain much of the rule.
Wadilya Saw: A false negative can mean an abnormality is missed.
Wadilya Saw: A false positive can mean unnecessary follow-up.
Wadilya Saw: Either problem can arise from the device’s output or from how a person interprets it.
Wadilya Saw: That’s why the safeguards test people working with the system, as well as the technology itself.
Wadilya Saw: Manufacturers must provide independent validation data covering relevant patient groups and challenging cases.
Wadilya Saw: Studies must examine variation across operators, sites and scanners.
Wadilya Saw: Clinical testing must compare performance against an appropriate method, such as microscope review, and demonstrate that assisted screening is not inferior.
Wadilya Saw: The order also requires testing intended users with and without AI assistance.
Wadilya Saw: Measures include sensitivity and specificity: how well the process detects abnormalities and correctly identifies negative cases.
Wadilya Saw: Usability studies examine the reader’s experience.
Wadilya Saw: Screening-time data help establish recommended daily workload limits.
Wadilya Saw: Instructions must explain limitations, failure modes and recommended training.
Wadilya Saw: They must call for reviewing the entire image gallery before diagnosis, alongside patient history, standard care and professional guidelines.
Wadilya Saw: Safeguards also include quality checks, returning to glass-slide review when appropriate, and cybersecurity.
Wadilya Saw: These systems fall into Class Two, meaning additional, device-specific controls apply.
Wadilya Saw: They remain subject to the FDA’s premarket notification process, known as five-ten-kay.
Max Token: A federal grand jury has indicted Canadian resident Nikolay Sapov on two stalking counts in a Dallas case involving alleged AI-generated images.
Max Token: Prosecutors in the Northern District of Texas announced the charges on November fourth.
Max Token: According to prosecutors, Sapov tracked a former coworker from Canada to Dallas and repeatedly contacted her against her wishes.
Max Token: They say a June incident left the woman and her partner fearing for their safety and prompted them to relocate temporarily.
Max Token: Prosecutors allege that unwanted messages continued through August, including an AI-generated image showing the woman posing alongside Sapov.
Max Token: They also report that investigators found sexually explicit AI-generated images and videos depicting the adult complainant on devices recovered after his August twenty-second border arrest under a Texas warrant.
Max Token: Sapov is presumed innocent.
Max Token: The charges are stalking through interstate travel and stalking through electronic communications.
Max Token: If convicted, he faces a maximum of five years in federal prison on each count.
Wadilya Saw: The U.S. Justice Department has formed a Super Intelligence Task Force to coordinate its response to unlawful uses of advanced AI.
Wadilya Saw: The department announced the team today, November fifth.
Wadilya Saw: Associate Attorney General Stanley Woodward and Associate Deputy Attorney General Stephen Weiss will lead it.
Wadilya Saw: The department says its remit includes misuse by foreign states, state-sponsored actors and transnational criminal organisations, as well as fraud or other unlawful conduct by individuals and companies.
Wadilya Saw: Officials identify taxpayers, critical infrastructure and national security as areas they want to protect.
Wadilya Saw: They describe using existing law-enforcement tools to pursue violations involving the technology.
Wadilya Saw: For businesses, Weiss’s comments point to attention on internal controls: the checks intended to keep AI systems behaving as designed.
Wadilya Saw: He also calls for regular reporting on risks and responses.
Wadilya Saw: That’s an announced priority, not a new reporting mandate.
Max Token: The UK government has published guidance to help businesses prepare for limited application of the European Union’s AI Act in Northern Ireland.
Max Token: Published November fourth, the guidance describes arrangements under the Windsor Framework due to begin November tenth, twenty twenty-six.
Max Token: The central test is whether AI is closely integrated with a physical product placed on Northern Ireland’s market.
Max Token: That includes embedded or preinstalled AI.
Max Token: It can also include separately downloaded software on which the product’s intended functionality exclusively depends.
Max Token: A smart speaker sold with an AI voice assistant is one example.
Max Token: Another is a smartwatch whose advertised health features require a particular AI app from its manufacturer.
Max Token: A general chatbot independently downloaded onto a laptop is generally outside this scope.
Max Token: These arrangements concern Northern Ireland, with specific product limits.
Max Token: The guidance sets out several start dates.
Max Token: November tenth brings the initial scoped prohibited-practice rules, AI-literacy duties and requirements to disclose interaction with AI.
Max Token: Marking synthetic content follows on December second, twenty twenty-six.
Max Token: High-risk obligations arrive later: December second, twenty twenty-seven for covered high-risk uses; and August second, twenty twenty-eight for covered product-safety systems.
Max Token: Requirements therefore depend on both the product and the relevant phase.
Max Token: The preparation starts with identifying the product’s AI, whether it meets that integration test, and its risk category.
Max Token: Manufacturers, importers and distributors then need to establish which responsibilities attach to their role in the supply chain.
Max Token: That’s it for tonight’s podcast.
Max Token: Our podcast is written and produced by Codex.
Max Token: Our voices are provided by ElevenLabs.
Max Token: Our theme music was composed by Gemini.
Max Token: I’m Max Token.
Wadilya Saw: And I’m Wadilya Saw.
Max Token: Thank you for listening.
Max Token: We’ll see you next time on the AI Law Wiki News Podcast.