Two reports examine how AI is being used in schools and newsrooms. In Maryland, Taki Allen is suing Baltimore County schools after a camera-based weapon detection system flagged a bag of Doritos as a gun. Allen says police handcuffed and searched him outside Kenwood High School after football practice, finding no weapon. The lawsuit, filed Nov. 17, alleges students’ constitutional rights were violated and that school officials and police ignored evidence. The district began using the system in 2023. Operator Omnilert expressed regret but said the alert led to rapid human verification, so its process worked as intended.
NPR, meanwhile, has launched Machine Gods, a podcast about AI whose hosts interview technology insiders and share their views. The hosts independently own and produce the show; NPR licenses it and sells sponsorships, with executives hoping to reach new audiences and sponsors. NPR’s newsroom journalists remain cautious about using AI in their work.
Questions about AI’s effects and limits also feature in research and public initiatives. Argentina’s Christian Alliance of Evangelical Churches (ACIERA) opened an Artificial Intelligence Observatory on Nov. 17 to study impacts on education, work, communication and relationships, guided by human dignity, rights and the common good. It will focus in part on children and teens, convene experts across disciplines, work with universities and policymakers, and provide churches with responsible-use training. Its agenda includes privacy, misinformation, child protection and healthy relationships.
A DeepMind researcher, Mr. Genewein, warns that artificial general intelligence could be followed by artificial superintelligence that outperforms people across many tasks. He says hardware improvements, more efficient systems and investment could increase AI computing power tenfold each year, while new training data, systems that generate ideas, collaboration among AI agents and self-improvement could accelerate progress. The article notes that critics question whether AGI is possible and that more computing power does not guarantee superintelligence. Genewein urges researchers, developers and policymakers to prepare and develop ways to measure progress. A separate analysis argues that AI’s gains may currently be more pronounced in working memory and attention span than intelligence: models can handle long, tedious tasks but still struggle to choose priorities or balance efficiency and perfection. AI may help research by connecting knowledge across fields, but real-world advances still require costly experiments, data and manufacturing. The article also describes research in which one AI model directly changed another’s weights to improve performance.
In health care, the FDA cleared A2z Radiology AI’s a2z-Omni-Triage for abdominopelvic CT scans, with or without contrast. It flags suspected cases of nine findings, including bowel perforation, fractures, abscesses, hydronephrosis and aortic aneurysms measuring at least 5 cm. The company says the tool can bring scans with suspected urgent findings forward for review.
AI is also changing how companies and universities work. Little Rock packaging-system maker GC Evans, which has served the food and beverage industry since the 1950s, uses AI to help engineers produce 3D floor plans and technical documents faster. President Mark McNamer says the tools let a small team offer services that once required more staff. AI cameras inspect up to 2,000 standard 12-ounce cans per minute for issues including damage, poor labels and faulty closures. McNamer says higher productivity could help smaller firms hire and compete with large international manufacturers.
At Politecnico di Milano, researchers in the META project studied responsibility, expertise and decisions involving agentic AI. Giacomo Zanotti found that greater autonomy can make it harder to identify who is responsible for a system’s actions. Camilla Quaresmini examined AI’s role in organizing information, making assessments and participating in research, including health care. Fabio Iapaolo’s work on self-driving cars found that their decisions emerge from interactions among sensors, algorithms, control devices and the surrounding environment.
In Armenia, ARCS∙ai opened the Armenian Robotics & Physical Intelligence center (ARPI), bringing robotics staff and equipment from universities together. Thirty students from the American University of Armenia and Yerevan State University are taking an undergraduate Physical AI minor with nine courses covering areas such as robot mechanics, reinforcement learning and digital twins. AUA plans to start a Physical AI master’s program in 2027. ARPI includes a new $500,000 robotics lab; AUA’s partnership is formalized, while a similar agreement with YSU remains in development.
Key Takeaways
- Taki Allen’s Nov. 17 lawsuit says a Baltimore County school AI system mistook Doritos for a gun, leading police to handcuff and search him; Omnilert says the alert prompted rapid human review.
- NPR’s Machine Gods is independently produced by its hosts and licensed by NPR, which sells sponsorships; newsroom staff remain cautious about AI use.
- ACIERA launched an Argentine AI observatory focused on human dignity, rights and the common good, including AI’s effects on children, privacy and misinformation.
- DeepMind researcher Mr. Genewein warns that AGI could be followed by ASI, while acknowledging that critics question AGI and that greater computing power alone does not ensure superintelligence.
- A separate analysis says today’s AI may be advancing working memory and attention more than intelligence, and that real-world research progress still depends on experiments, data and manufacturing.
- The FDA cleared A2z Radiology AI’s a2z-Omni-Triage to flag nine suspected findings on abdominopelvic CT scans, including aortic aneurysms of at least 5 cm.
- GC Evans uses AI for engineering documents and 3D floor plans; AI cameras inspect up to 2,000 standard 12-ounce cans per minute. President Mark McNamer says productivity could help the small manufacturer compete and hire.
- Politecnico di Milano’s META research examines how agentic AI complicates responsibility, supports knowledge work and makes decisions through interactions with its physical environment.
- Armenia’s ARPI center opened a $500,000 robotics lab and supports a nine-course Physical AI minor for 30 students; AUA plans a master’s program in 2027.
Maryland student sues after AI system flags Doritos as a gun
Taki Allen was sitting with friends outside Kenwood High School in Essex, Maryland, after football practice when armed police arrived and handcuffed him. Allen said officers searched him and found he had nothing. A lawsuit filed Nov. 17 in Baltimore County Circuit Court says the students’ constitutional rights were violated after a security alert and after school officials and police ignored evidence. Baltimore County schools began using the AI system in 2023, and it uses existing cameras to detect weapons. Its operator, Omnilert, said the process worked as intended because it led to quick human review, even though the object was not a firearm.
Maryland student sues after AI system flags Doritos as a gun
Taki Allen was sitting with friends outside Kenwood High School in Essex, Maryland, after football practice when armed police arrived and handcuffed him. Allen told WBAL that officers searched him and found he had nothing. A lawsuit filed Nov. 17 in Baltimore County Circuit Court alleges that students’ constitutional rights were violated by the technology and by school officials and police who disregarded evidence. Baltimore County public schools began using the camera-based AI weapon detection system in 2023, according to Superintendent Myriam Rogers. Omnilert, which runs the system, expressed regret but said the process worked as intended by prompting rapid human verification.
NPR’s AI podcast uses a new business model as newsroom remains cautious
NPR debuted Machine Gods, a podcast about AI, shortly before the midterms. Its hosts interview technology insiders and discuss AI developments, while sharing personal views. The hosts independently own and produce the show, while NPR licenses it and sells sponsorships. NPR executives see the podcast as a chance to attract new sponsors and audiences. At the same time, NPR journalists are taking a cautious approach to using AI in their newsroom work.
Argentine evangelical alliance opens observatory on AI’s effects on people
The Christian Alliance of Evangelical Churches of the Argentine Republic, known as ACIERA, launched its Artificial Intelligence Observatory on Nov. 17. The group will study how AI affects education, work, communication and human relationships, with human dignity, rights and the common good as guiding principles. Director Ayelén Caffarena said one focus will be AI’s effects on children and teens, including their learning, relationships and development. The observatory will bring together experts from fields such as law, technology, neuroscience, education, theology and ethics, and plans to work with universities, civil society groups and policymakers. It will also offer churches training on responsible AI use and address privacy, misinformation, child protection and healthy relationships.
DeepMind warns AI could move beyond human-level abilities
DeepMind researcher Mr. Genewein says artificial general intelligence, or AGI, could be followed by artificial superintelligence, or ASI, which would outperform many people across a wide range of tasks. He argues that hardware advances, more efficient systems and major investment could increase the computing power available for AI work tenfold each year. New training data, systems that can develop ideas, better teamwork among AI agents and self-improvement could also speed progress. The article notes that some critics doubt AGI is possible and that greater computing power does not guarantee ASI. Genewein says researchers, developers and policymakers should prepare for what may follow human-level AI and build ways to measure progress.
FDA Clears AI Tool to Flag Nine Findings on Abdominopelvic CT
The FDA has cleared a2z-Omni-Triage, an AI-powered tool from A2z Radiology AI, for use with abdominopelvic CT scans. It flags suspected cases of nine findings, including bowel perforation, fractures, abscesses, hydronephrosis and large aortic aneurysms measuring 5 cm or more. The software can be used with contrast-enhanced and non-contrast CT scans. Michael Moritz, chief medical information officer at A2z Radiology AI, said the tool helps bring scans with suspected urgent findings forward for review.
AI May Be Expanding Memory and Attention More Than Intelligence
The article argues that current AI may be experiencing an explosion in working memory and attention span, rather than a sudden rise in intelligence. The writer says frontier models can complete long, tedious tasks but still struggle to choose where to focus or balance efficiency with perfection. AI could speed up progress in research by combining knowledge across fields, but real-world advances also depend on costly experiments, data and manufacturing. The article also notes research in which one AI model directly changed another model’s weights to improve its performance.
AI Helps Arkansas Manufacturer Compete With Global Companies
Little Rock based GC Evans has made automated packaging systems for the food and beverage industry since the 1950s. President Mark McNamer says AI tools help its engineers create 3D floor plans and technical documents faster, allowing a small team to offer services that once required more staff. AI cameras can inspect packaging at speeds of up to 2,000 standard 12 ounce cans per minute and identify problems such as damaged cans, poor labels and faulty closures. McNamer said the added productivity could help smaller companies like GC Evans hire more workers and compete with large international manufacturers.
Three Studies Explore AI Responsibility, Expertise and Real World Decisions
Researchers at Politecnico di Milano’s META project studied how agentic AI affects responsibility, knowledge and decisions in the physical world. Giacomo Zanotti’s research found that greater AI autonomy can make it harder to determine who is responsible for a system’s actions. Camilla Quaresmini examined how AI can organize information, make assessments or take part in research, including in health care. Fabio Iapaolo studied self-driving cars and found that their decisions arise through interactions among sensors, algorithms, control devices and the surrounding environment.
Armenian Universities Launch Shared Center for Physical AI Training
ARCS∙ai has opened the Armenian Robotics & Physical Intelligence center, known as ARPI, to combine robotics staff and equipment from Armenian universities. Thirty students from the American University of Armenia and Yerevan State University are enrolled in an undergraduate Physical AI minor, with nine courses covering topics such as robot mechanics, reinforcement learning and digital twins. AUA plans to launch a Physical AI master’s program in 2027. The center includes a new $500,000 robotics lab, while the AUA partnership is formalized and a similar agreement with YSU remains in development.
Sources
- Maryland students sue after AI security mistook bag of Doritos for gun
- Maryland students sue after AI security mistook bag of Doritos for gun
- 'Machine Gods,' NPR's new podcast about AI, is a unique business deal
- Argentine evangelical alliance launches observatory to study AI’s impact on human life
- Google says AI is about to pass a huge new threshold – and we need to prepare
- FDA Clears Emerging AI Triage Software for Nine Findings on Abdominopelvic CT
- Does it really matter if there’s an intelligence explosion?
- Last Word: How Artificial Intelligence Is Making Arkansas Business Internationally Competitive
- Three studies examine how autonomous AI reshapes responsibility, expertise and real-world decisions
- ARCS∙ai Opens Shared ARPI Center for Physical AI Training
Comments
Please log in to post a comment.