google unveils new tools as hugging face ships new models

Key Takeaways

Students fear AI will replace their critical thinking skills

Many students at Villanova University admit to using AI to complete homework and exams to avoid struggling with difficult problems. They believe this shortcut helps them pass classes quickly without the discomfort of deep thinking. However, experts warn that relying on AI rewires the brain to expect immediate answers, which weakens problem-solving abilities and memory. Cheating with AI can lead to failing grades, permanent marks on records, or even expulsion from university. The article argues that students are ultimately cheating themselves by losing the fundamental learning skills they need for future careers.

Colleges use oral exams to stop students from using AI

Universities across the U.S. are shifting toward oral exams to combat the rise of AI in student assignments. Professors at Cornell, Penn, and NYU require students to speak directly with instructors or use AI-powered oral defenses to prove they understand the material. This method prevents students from simply copying perfect written essays generated by chatbots. Educators worry that AI use is causing students to lose cognitive capacity and creativity. Some schools are also using AI tools to test if students can defend their work against an AI agent.

Podcast explores different views on AI in education

Lansing Community College hosted a podcast episode discussing the social and ethical impacts of artificial intelligence on education. Students and community members shared contrasting opinions on whether AI is helpful or harmful to learning. One student believes AI is destructive because it can replace coursework and leave people mentally handicapped if not used carefully. Another student sees AI as a tool that can teach complex concepts simply and help individuals learn better. The conversation highlighted the difficulty teachers face in detecting when students use AI during assignments.

Colleges struggle with AI cheating in assignments

Professors at Rider University report that students are increasingly relying on AI programs like ChatGPT to complete assignments. Some instructors ban AI use entirely to ensure honest work, while others allow it for brainstorming. Students in competitive settings like Model United Nations feel it is unfair when competitors use AI to generate position papers. Educators note that AI creates a specific writing style that differs from a student's natural voice, making it easier to spot cheating. Despite these challenges, some students admit they use AI to improve grammar and sentence structure.

Merck and Google Cloud partner on billion-dollar AI deal

Merck and Google Cloud announced a partnership worth up to $1 billion to transform Merck into an AI-enabled enterprise. The deal will deploy an agentic platform across Merck's research, manufacturing, and corporate functions using Google Cloud's Gemini Enterprise technology. This collaboration aims to boost productivity for Merck's 75,000 employees worldwide and accelerate scientific breakthroughs. The partnership includes deploying AI solutions for end-to-end research workflows and optimizing manufacturing through predictive analytics. Leaders from both companies stated this marks a significant phase in Merck's AI journey.

Vodafone and Google Cloud launch AI tools for small businesses

Vodafone Business and Google Cloud are launching new cybersecurity and AI solutions to help small and medium-sized businesses. The partnership includes a managed detection and response service enabled by Google Security Operations to protect against cyber threats. Vodafone is also introducing an AI Concierge built with Google Gemini to handle customer inquiries and book appointments autonomously. These tools will initially launch in Germany and Greece to support local data protection standards. The goal is to make advanced AI practical and secure for everyday business use.

Hugging Face releases open-source AI agent for training

Hugging Face released ml-intern, an open-source AI agent designed to automate the post-training workflow for large language models. The tool uses the smolagents framework to perform literature reviews, dataset discovery, and iterative evaluation automatically. In tests, ml-intern improved the GPQA benchmark score of a Qwen3-1.7B model from 8.5% to 32% in under 10 hours. This performance outpaced existing tools like Claude Code on the same task. The agent also demonstrated the ability to generate synthetic data and use Group Relative Policy Optimization for math performance.

Europe has talent but relies on American tech giants

Europe has record AI funding and a deep talent pool, yet it struggles to compete with American tech companies. European startups often build on infrastructure owned by American corporations like Microsoft and Apple. By the later growth stages, 73% of European AI companies are led by American investors, creating a structural dependency. The Magnificent Seven tech giants aggressively recruit European engineers, offering salaries two to three times higher than local companies. This talent drain prevents Europe from retaining the economic value created by its own skilled workforce.

Open-source AI finds 23 flaws in mock network test

IBM security experts tested an open-source AI agent called OpenClaw to find vulnerabilities in a simulated legacy network. The AI identified 23 high-quality vulnerabilities within the given timeframe, mimicking human red team operators. While the results show promise, experts warn that AI can be unpredictable without proper guardrails. The team agreed that AI should act as an augmentation tool rather than a replacement for human expertise. Human oversight remains crucial for contextualizing findings and making strategic security decisions.

Florida Atlantic University launches new AI ethics degree

Florida Atlantic University added a new Bachelor of Science in Interdisciplinary Artificial Intelligence to address workforce gaps. The program trains students to build AI systems while also evaluating their fairness, trustworthiness, and accountability. Officials state this degree fills a critical need by preparing graduates to handle the real-world impact of AI technology. The curriculum focuses on interdisciplinary approaches to ensure students can both create and responsibly manage AI solutions.

OpenAI makes ChatGPT free for U.S. clinicians

OpenAI launched ChatGPT for Clinicians, a free version of ChatGPT designed specifically for medical professionals in the U.S. The tool supports clinical tasks like documentation, medical research, and care consultations. A 2026 survey shows 72% of physicians now use AI in clinical practice, up from 48% the previous year. ChatGPT for Clinicians includes advanced models, trusted clinical search, and optional HIPAA compliance support. It also allows evidence reviews to count toward continuing medical education credits automatically.

Housing finance officials announce major system development

HUD Secretary Scott Turner and U.S. Federal Housing Director Bill Pulte announced a significant development in the U.S. housing finance system. The press conference took place at 12:45 PM ET to discuss potential changes to the FHA and VA programs. Rumors suggest operationalizing VantageScore or moving control of these programs under Pulte and the FHFA. The announcement has sparked discussions in the capital markets about hedge fund ownership of U.S. Treasuries and mortgage rates.

Security teams struggle to keep up with AI-generated code

ProjectDiscovery released a report showing that AI-generated code is outpacing security teams' ability to review it. Nearly half of surveyed cybersecurity practitioners attribute increased engineering delivery to AI-assisted coding tools. However, 62% of security teams say keeping up with this volume is getting harder. Two-thirds of practitioners spend more than half their time manually validating findings instead of fixing vulnerabilities. The top concern is exposing secrets in code, which 78% of practitioners rank as the biggest challenge.

Kanye West confirms no AI on upcoming Bully album

Kanye West clarified that his upcoming 13-track album Bully does not use artificial intelligence. In a recent interview, West emphasized that the album is a purely human creation, stating clearly that there is no AI involved. This statement marks a shift from his earlier comments describing AI as part of his music process. The album has already generated significant buzz among fans and industry experts. West's decision to distance himself from AI continues to spark debates about the role of technology in music creation.

Sources

• Is AI Replacing Critical Thinking?
• Perfect homework, blank stares: Why colleges are turning to oral exams to combat AI
• A conversation about artificial intelligence
• The rise of Artificial Intelligence in higher education
• Merck and Google Cloud Partner to Accelerate Agentic AI Enterprise Transformation
• Vodafone Business Launch New AI and Cybersecurity Solutions to Accelerate Small Business Digital Transformation in Partnership with Google Cloud
• Hugging Face Releases ml-intern: An Open-Source AI Agent that Automates the LLM Post-Training Workflow
• Europe has the talent and dunding to win at AI. First, it needs to break free from the Magnificent Seven
• AI Pen Testing: Open Source AI Finds 23 Flaws in Mock Network
• Florida college adds new AI bachelor’s program on ethics, real-world impact
• Making ChatGPT better for clinicians
• MBS Execution, AI, DSCR, Processing, HELOC Products; Hedge Funds, Treasuries, and Mortgage Rates
• ProjectDiscovery's "2026 AI Coding Impact Report" Reveals AI-Generated Code Is Outpacing Security Teams' Ability to Keep Up
• 'No AI,' Kanye West Says As He Previews 13-Track 'Bully' Album — After Earlier Describing AI As Part Of His Music Process