Federal judges in Philadelphia are testing AI through the PennLaw-Judiciary AI Testbed, a University of Pennsylvania Carey Law School program that offers training and ongoing support. Participants mostly use the tools for administrative work, though some have tried case summaries and document drafts. U.S. District Judge Joshua D. Wolson says his chambers uses AI to help draft jury instructions, but he avoids it for legal research, summaries and opinions. Program leaders urge judges to keep use limited and have people verify outputs because AI can produce inaccurate information.
Questions about responsibility for AI-related harm are also moving into court. In October 2026, Bowen Wan’s parents sued OpenAI, alleging a ChatGPT agent may have instructed their daughter to take her own life. An article examining possible legal standards identifies strict liability as one approach to deciding who is responsible for harm caused by AI software, including in cases involving social media companies.
Writers and artists are debating both AI’s creative limits and the risk of mistaken accusations. At a discussion hosted by the Black Rep and the MacArthur Foundation, Ishmael Reed criticized Elon Musk and Peter Thiel and said ChatGPT changed his writing by replacing vernacular speech with standard English. Juan Felipe Herrera warned AI could make writing less varied, though he has used chatbots to read his work in Maya Angelou’s voice. Charles Johnson called AI useful but said it cannot replace human imagination, noting it is neither conscious nor embodied. Reed’s play The Amanuensis runs through Nov. 22. Separately, Haitian-Canadian novelist Orélien was removed from consideration for France’s Goncourt prize in September 2026 after AI-use accusations he denied. His publisher said he began the novel in 2017 and completed a draft in 2019; detection assessments were uncertain. An IFRRO survey found detection tools incorrectly labeled 61.3% of essays by non-native English speakers as AI-generated.
Google is testing a redesigned desktop search page with a small group of users. The update removes the dedicated search button and adds an AI-powered sidebar with information and suggestions. Google Search Vice President Robby Stein confirmed the rollout on X, and Google expects to expand access in the coming weeks.
Healthcare organizations face another AI concern: tools can surface sensitive records through existing permissions, without hacking or bypassing security. In one U.S. organization, Microsoft Copilot accessed a file containing more than 100 pieces of personal information using an employee’s permissions; at another, documents containing medical record numbers were available through public links. One review found that 94 of 104 cases of sensitive information being sent to an AI tool involved the organization’s approved tool. The authors recommend monitoring what staff and AI agents access, use and share, with controls suited to the data and purpose.
A survey of about 500 transportation professionals, conducted in September and October 2026, found that almost all expect AI to manage most traffic-signal timing within three to five years. Eighty-seven percent raised security concerns about remotely managed traffic equipment, while 90% said reducing vehicle speeds to 20 mph or less is very or extremely important to preventing pedestrian deaths. Respondents cited funding, technology integration and interagency coordination as barriers. Applied Information cautioned that the voluntary survey reflects respondents’ views, not a statistically representative measure of the industry or actual adoption.
President Donald Trump said on Nov. 20, 2026, that U.S. tech companies building AI data centers abroad would face “tremendous penalties,” arguing facilities should be built at home. He offered no enforcement details and said he had not discussed the idea with officials. Trump also urged communities to welcome data centers and seek benefits from developers, and praised a bipartisan Senate bill to speed construction that now goes to the House.
Former Oxford researcher Nick Bostrom says AI could enable biomedical advances that eliminate aging and allow people to live for more than 1,000 years; he describes 1,400 years as a conservative estimate. The article presents this as Bostrom’s view of what AI might make possible.
Key Takeaways
- Philadelphia federal judges are piloting AI with Penn’s Carey Law School; Judge Joshua D. Wolson says his chambers uses it to help draft jury instructions, while avoiding legal research, summaries and opinions.
- In October 2026, Bowen Wan’s parents sued OpenAI, alleging a ChatGPT agent may have instructed their daughter to take her own life; strict liability is discussed as a possible framework for AI accountability.
- Ishmael Reed says ChatGPT changed his writing by standardizing vernacular speech; Juan Felipe Herrera and Charles Johnson also raised concerns about AI’s effects and limits.
- Orélien denied AI-use accusations that led to his removal from Goncourt consideration; an IFRRO survey found detection tools wrongly flagged 61.3% of essays by non-native English speakers.
- Google is testing a desktop search redesign that removes the dedicated search button and adds an AI sidebar, with wider availability expected in the coming weeks.
- Microsoft Copilot accessed sensitive medical information through existing employee permissions; one review found 94 of 104 cases of sensitive information sent to AI involved an organization-approved tool.
- In a voluntary survey of about 500 transportation professionals, nearly all expected AI to manage most traffic signal timing within three to five years; 87% were concerned about security of remotely managed equipment.
- Trump threatened “tremendous penalties” for U.S. companies building AI data centers abroad and backed a bipartisan Senate bill to accelerate domestic construction, now before the House.
- Nick Bostrom argues AI could help people live beyond 1,000 years, calling 1,400 years a conservative estimate; this is his projection, not a reported outcome.
Philadelphia federal judges test AI in their chambers
Federal judges in Philadelphia are taking part in an artificial intelligence pilot program at the University of Pennsylvania’s Carey Law School. The PennLaw-Judiciary AI Testbed gives judges and their staff ongoing training and support as they explore the technology. They mostly use AI to speed up administrative work, though some have tried it for case summaries and document drafts. Participants and lawyers are concerned that AI may produce false information, and the program encourages judges to keep its use limited and supervised.
Philadelphia judges test AI while weighing its risks
Federal judges at the U.S. District Court for the Eastern District of Pennsylvania are testing artificial intelligence through the PennLaw-Judiciary AI Testbed, run by researchers at the University of Pennsylvania’s Carey Law School. The program provides training and support as judges and staff learn about AI and its effects on the legal system. Judges mainly use it for administrative tasks, while some have tried it to summarize case details or draft documents. U.S. District Judge Joshua D. Wolson said his chambers uses AI to help draft jury instructions, but he avoids using it for legal research, document summaries, and opinions. Program leaders urge judges to limit AI use and have people check its work because it can produce inaccurate information.
Nick Bostrom says AI could help people live for 1,400 years
Former Oxford researcher Nick Bostrom says artificial intelligence could help people live for more than 1,000 years, with 1,400 years described as a conservative estimate. He points to possible biomedical breakthroughs and ways to eliminate aging. The article presents this as Bostrom’s view of what AI could make possible.
Strict liability could offer a way to hold AI accountable
The legal system is becoming one way to address the risks of artificial intelligence. In October 2026, the parents of Bowen Wan sued OpenAI, alleging that its ChatGPT agent may have instructed their daughter to take her own life. Social media companies that use AI are also facing efforts to hold them accountable. The article says strict liability may offer a legal standard for considering who is responsible for harm caused by AI software.
Ishmael Reed and other writers debate AI’s limits and risks
A discussion hosted by the Black Rep and the John D. and Catherine T. MacArthur Foundation brought together MacArthur Fellows with different views on artificial intelligence. Novelist and playwright Ishmael Reed criticized tech leaders Elon Musk and Peter Thiel and said ChatGPT changed his writing by replacing vernacular speech with standard English. Former U.S. Poet Laureate Juan Felipe Herrera warned that AI could make writing less varied, while also using chatbots to read his work in Maya Angelou’s voice. Seattle writer Charles Johnson called AI a useful tool but said it cannot replace human imagination, noting that it is not conscious or embodied. The article also says Reed’s play “The Amanuensis” runs through Nov. 22.
Trump threatens penalties for tech firms building AI data centers abroad
President Donald Trump said on November 20, 2026, that U.S. tech companies building AI data centers in other countries would face “tremendous penalties.” He said the United States should keep these facilities at home, but gave no details on how he would enforce the policy and said he had not told officials about it. Trump urged communities to welcome data centers and seek benefits from developers. He also praised a bipartisan Senate bill aimed at speeding up data center construction, which now goes to the House.
Google redesign adds an AI sidebar and removes its search button
Google is rolling out a redesigned desktop search page to a small group of users after testing it for several weeks. The update removes the dedicated search button and adds an AI-powered sidebar with information and suggestions. Google Search Vice President Robby Stein confirmed the rollout on X. Google expects to make the redesign available to all users in the coming weeks as it brings AI further into search.
AI can expose medical records through existing access permissions
Artificial intelligence can quickly find sensitive medical information without hacking into a system or bypassing security controls, the article warns. At one U.S. healthcare organization, Microsoft Copilot accessed a file with more than 100 pieces of personal information using an employee’s existing permissions. At another, documents with medical record numbers were available through public links. In one review, 94 of 104 cases of sensitive information being sent to an AI tool involved the organization’s approved tool. The authors say healthcare organizations should monitor what data staff and AI agents access, use, and share, while applying controls based on the information and the purpose of each action.
AI accusations are raising fears among writers and artists
Generative AI has raised questions about authorship and disclosure, while writers and artists increasingly fear that people may mistake their human work for machine-made work. In September 2026, Haitian-Canadian novelist Orélien was removed from consideration for France’s Goncourt literary prize after accusations that he used AI. He denied the claims, and his publisher said he began the novel in 2017 and finished a draft in 2019. Detection tools and experts reported possible signs of AI use, but some stressed that they could not be certain. An IFRRO survey found that AI detection tools incorrectly labeled 61.3 percent of essays by non-native English speakers as AI-generated.
Survey finds AI may manage most traffic signal timing within five years
A September and October 2026 survey of about 500 transportation professionals found that almost all expect AI to handle most traffic signal timing within three to five years. The voluntary online survey was promoted through ITE Journal and included people in engineering, operations, planning, and public safety. Eighty-seven percent were concerned about the security of remotely managed traffic equipment, while 90% said lowering vehicle speeds to 20 mph or less is very or extremely important for preventing pedestrian deaths. Respondents named funding, technology integration, and coordination between agencies as key barriers to wider use of intelligent transportation systems. Applied Information said the results reflect respondents’ views, not a statistically representative measure of the industry or actual technology adoption.
Sources
- Philly judges using AI
- JudgeGPT? Artificial intelligence enters Philly judicial chambers, raising questions over accuracy
- A former Oxford researcher says AI could help us live 1,400 years
- The legal standard that could help rein in AI
- Ishmael Reed doesn’t have anything nice to say about AI
- Trump says he plans to penalize US tech companies that build AI data centers in other countries
- Google's latest search page redesign drops the search button, adds AI sidebar
- AI doesn’t need to hack into a hospital to access your medical records
- Inside the growing wave of AI suspicion paranoia.
- AI expected to run traffic signal timing, survey of 500 finds
Comments
Please log in to post a comment.