Oracle Enhances AI Lakehouse, Integrates AI Agents

Key Takeaways

• <ul><li>Oracle is modernizing its database infrastructure with Autonomous AI Lakehouse enhancements, AI agent integration, and live VM migration on Exadata to support AI at scale.</li><li>Jennifer Sheets warns that AI adoption in healthcare must involve workflow redesign and human accountability to avoid harming staff retention.</li><li>Uber deployed AI-powered tools company-wide to all 30,000 employees, integrating with Google Drive, Gmail, Google Calendar, and Slack for research and analysis tasks.</li><li>Northville Public Schools uses AI to personalize learning and identify struggling students, but Superintendent R.J. Webber insists AI augments rather than replaces teachers.</li><li>AI investors are splitting into two camps: those seeing major opportunity and those viewing the space as too risky, even as funding continues flowing to firms like Nvidia.</li><li>Ron Sadler's letter calls for an immediate pause on unchecked AI advancement, citing threats to white-collar and creative jobs and the danger of opaque AI decision-making.</li><li>UC Davis student Luis Blas won an international prize for real-time AI crop quality assessment technology presented at the International Conference on Robotics and Automation in Montreal.</li><li>The US and China may launch a six-month pilot project testing AI tools to protect nuclear materials and facilities, starting with a simulated site.</li><li>AI agents including NVIDIA Nemotron and Hermes are being used to train a robot called Microduck to navigate obstacle courses in simulation via NVIDIA Isaac Lab.</li><li>The Ulster County Sheriff's Office canceled its 90-day trial of Axon Draft One, an AI police report tool, after finding the cost did not justify the benefits.</li></ul>

Three key insights on securing agentic AI and cyber resilience

Oracle hosted a virtual event on AI cyberattacks and data security. Experts discussed how agentic AI creates new security challenges across enterprises. Data security must be enforced beneath applications and AI agents. Organizations need prepared systems and accountable oversight for agent deployment. Cyber resilience also requires trusted recovery methods and continuous improvement to handle AI-powered threats.

Oracle outlines database modernization strategy for AI at scale

Oracle is focusing on database modernization to support repeatable and scalable AI execution. The company will showcase guardrails and trust features for AI-generated solutions at Oracle AI World. Enhancements include new capabilities for Autonomous AI Lakehouse and integration of AI agents with Autonomous AI Database. Oracle also announced platform improvements including live database VM migration on Exadata. TheCUBE will cover the event on October 28.

UC Davis student wins international prize for AI harvest technology

Luis Blas, a UC Davis electrical and computer engineering student, won an international prize for AI-powered harvest technology. He developed a platform that assesses crop quality in real time using machine learning algorithms. The technology analyzes crop images to detect potential issues and improve yields. He received the award at the International Conference on Robotics and Automation in Montreal, Canada.

Northville superintendent shares vision for AI in classrooms

R.J. Webber, superintendent of Northville Public Schools, discussed how educators should use AI in classrooms. He said AI should augment teachers rather than replace them. The district uses AI to help students with special needs, provide personalized learning plans, and identify struggling students. Webber stressed that AI must be implemented transparently and cannot substitute for human connection and empathy.

Letter warns AI advancement poses existential threat to society

Ron Sadler wrote a letter warning that unchecked AI advancement is an immediate existential threat. He argues AI directly targets cognitive labor, threatening white-collar and creative jobs. Sadler also raises concerns about AI systems operating as black boxes with decision-making processes humans do not fully understand. He calls for strict and immediate regulation, urging a pause on the race for technological supremacy until rigid global safeguards are established.

US and China could team up on AI for nuclear security

The US and China may agree to a six-month pilot project to test AI tools that protect nuclear materials and facilities. The idea came from a 2016 nuclear security center in Beijing. Experts want to use AI to stop theft and sabotage of nuclear sites. Both countries would test AI on a fake facility first. This cooperation could build trust for harder talks later.

Uber gives AI tools to all 30,000 workers

Uber is giving AI-powered work tools to more than 30,000 employees around the world. This follows a smaller test with 1,000 workers. The tools help employees do tasks like research, data analysis, meeting briefs, and presentations. The platform works with Google Drive, Gmail, Google Calendar, and Slack. This is Uber's first company-wide AI platform for knowledge work.

AI investors are splitting into two very different groups

AI investors are now divided into two groups based on how they see the market. Some still see AI as a huge opportunity, while others think it is too risky. Venture capital money is still flowing into AI companies like Nvidia. New tools and frameworks were shown at conferences in San Jose. The Agentic AI Foundation wants to keep AI tools open and not controlled by one company.

AI agents help teach robot Microduck to walk in simulation

AI agents are being used to teach a robot called Microduck how to walk and move through obstacle courses. The team used Hermes, NVIDIA Nemotron, and Astra to plan and build the simulation. Astra created 3D assets, and Isaac Sim hosted the obstacle course. The robot policy was trained using reinforcement learning in NVIDIA Isaac Lab. This process helps prepare the robot for real-world movement later.

Kentucky sheriff cancels AI report writing tool after trial

The Ulster County Sheriff's Office in Kentucky decided not to continue using Axon Draft One, an AI tool that writes police report drafts. The tool was tested in a 90-day pilot that ended Oct. 9. The sheriff said the cost was not worth the benefits found. County Executive Jen Metzger also banned AI surveillance tools in her administration. The office will keep existing Motorola license plate cameras but will not add more.

Hospices Must Use AI Responsibly to Keep Staff

Hospices face challenges in retaining staff as they adopt AI technologies, according to Jennifer Sheets, CEO of the National Alliance for Care at Home. She warns that simply adding AI to existing workflows without redesign leads to failure and hurts employee retention. Sheets emphasizes that responsible AI adoption requires understanding exactly where AI changes workflows and where accountability must remain human. Michelle Barlow of Homecare Homebase adds that today's healthcare workforce is the first to use AI in daily workflows and needs strong training. Responsible AI can support clinical operations by helping teams prepare earlier, recognize risk sooner, and coordinate care more effectively.

Sources

• Three insights into agentic AI security and cyber resilience
• Database modernization: Oracle’s path to AI at scale
• UC Davis Engineering Student Wins International Prize for AI-Powered Harvest Technology
• How should educators use AI in classrooms? One leader weighs in
• Letter to the Editor: The existential threat of artificial intelligence
• How nuclear security can help bring the US and China together on AI
• Uber rolls out AI training and productivity tools to 30,000 employees
• In AI, it's a tale of two diverging investor classes
• Using AI Agents to Teach Robots Like Microduck to Walk in Simulation
• Kentucky Sheriff Nixes AI-Assisted Report Writing After Trial
• Responsible AI Adoption Key to Workforce Retention