All news with #ai safety tag

🎧 In episode 74 of The AI Fix, hosts Graham Cluley and Mark Stockley survey recent AI developments including Amazon’s experimental delivery glasses, Channel 4’s AI presenter, and reports of LLM “brain rot.” They examine practical security risks — such as malicious browser extensions spoofing AI sidebars and AI browsers being tricked into purchases — alongside wider societal debates. The episode also highlights public calls to pause work on super-intelligence and explores what AGI really means.

⚠️ Bruce Schneier highlights how facial recognition systems frequently fail people with nonstandard facial features, producing concrete barriers to services and daily technologies. Those interviewed report being denied access to public and financial services and encountering nonfunctional phone unlocking and social media filters. The author argues the root cause is often design choices by engineers who trained models on a narrow range of faces and calls for inclusive design plus accessible backup systems when biometric methods fail.

⚠️A new ISACA report identifies AI-driven social engineering as the top cyber threat for 2026, cited by 63% of nearly 3,000 IT and security professionals. The 2026 Tech Trends and Priorities report, published 20 October 2025, shows AI concerns outpacing ransomware (54%) and supply chain attacks (35%), while only 13% of organizations feel very prepared to manage generative AI risks. ISACA urges organizations to adopt AI governance, strengthen compliance amid divergent US and EU approaches, and invest in talent, resilience and legacy modernization.

📱 Stephen Balkam of FOSI argues that instead of blanket bans, families benefit from thoughtful restrictions, ongoing dialogue and tools that preserve teen agency. He highlights solutions such as Family Link and YouTube’s supervised experience and proposes that AI assistants (for example, Gemini or ChatGPT) could configure age-, app- and device-specific controls. He urges coordinated action from policymakers, teachers and parents and calls for impartial digital literacy and AI education frameworks.

🔒 Google for Education highlights built-in security, responsible AI, and partnerships to create safer digital learning environments for schools, classrooms, and families. Admins benefit from automated 24/7 monitoring, encryption, spam filtering, and security alerts; Google reports zero successful ransomware attacks on Chromebooks to date. Gemini for Education and NotebookLM provide enterprise-grade data protections with admin controls and age-specific safeguards, while family resources and a $25M Google.org clinics fund extend protection and workforce development.

⚠️A new primer, Scam GPT, surveys how generative AI is being adopted by criminals to automate, scale, and personalize scams. It maps which communities are most at risk and explains how broader economic and cultural shifts — from precarious employment to increased willingness to take risks — amplify vulnerability to deception. The author argues these threats are social as much as technical, requiring cultural shifts, corporate interventions, and effective legislation to defend against them.

🔍 In episode 70 of The AI Fix, hosts Graham Cluley and Mark Stockley examine how AI alters human behaviour and how deployed systems can fail in unexpected ways. They discuss research showing AI can increase dishonest behaviour, Waymo's safety record and a mirror-based trick that fooled self-driving perception, a rescue robot that mishandles victims, and a Chinese fusion-plant robot arm with extreme lifting capability. The show also covers a demonstration of a ChatGPT agent solving image CAPTCHAs by simulating mouse movements and a paper on deliberative alignment that functions until the model realises it is being watched.

🔒 Microsoft and xAI have brought Grok 4 to Azure AI Foundry, combining a 128K-token context window, native tool use, and integrated web search with enterprise safety controls and compliance checks. The release highlights first-principles reasoning and enhanced problem solving across STEM and humanities tasks, plus variants optimized for reasoning, speed, and code. Azure AI Content Safety is enabled by default and Microsoft publishes a model card with safety and evaluation details. Pricing and deployment tiers are available through Azure.

🔒 OpenAI confirmed that when GPT-4o detects sensitive, emotional, or potentially harmful activity it may route individual messages to a dedicated safety model, reported by some users as gpt-5-chat-safety. The switch occurs on a per-message, temporary basis and ChatGPT will indicate which model is active if asked. The routing is implemented as an irreversible part of the service's safety architecture and cannot be turned off by users; OpenAI says this helps strengthen safeguards and learn from real-world use before wider rollouts.

🤖 Microsoft is testing a new AI-powered Auto-Categorization capability in Microsoft Photos on Windows 11, rolling out to Copilot+ PCs across all Windows Insider channels. The feature automatically groups images into predefined folders — screenshots, receipts, identity documents, and notes — using a language-agnostic model that recognizes document types regardless of image language. Users can locate categorized items via the left navigation pane or Search bar, manually reassign categories, and submit feedback to improve accuracy. Microsoft has not yet clarified whether image processing happens locally or is sent to its servers.

🧬 Escalante leverages JAX's functional, composable design to combine many predictive models into a single differentiable objective for protein engineering. By translating models (including AlphaFold and Boltz-2) into a JAX-native stack and composing them serially or linearly, they compute gradients with respect to input sequences and evolve candidates via optimization. Each job samples thousands of sequences, filters to roughly ten lab-ready designs, and runs at scale on Google Kubernetes Engine using spot TPU v6e, yielding a reported 3.65x performance-per-dollar advantage over H100 GPUs.

🤖 The 2025 DORA Report synthesizes survey responses from nearly 5,000 technology professionals and over 100 hours of qualitative data to examine how AI is reshaping software development. It finds AI amplifies existing team strengths and weaknesses: strong teams accelerate productivity and product performance, while weaker teams see magnified problems and increased instability. The report highlights near-universal AI adoption (90%), widespread productivity gains (>80%), a continuing trust gap in AI-generated code (~30% distrust), and recommends investment in platform engineering, user-centric workflows, and the DORA AI Capabilities Model to unlock AI’s value.

🛡️ IT security is entering a new era where autonomy augments human defenders, moving beyond scripted automation to adaptive, AI-driven responses. Traditional playbooks and scripts are limited because they only follow defined rules, while attackers continuously change tactics. Organizations must adopt self-driving security systems that combine real-time telemetry, machine learning, and human oversight to improve detection, reduce response time, and manage risk.

🎧 In episode 68 of The AI Fix, hosts Graham Cluley and Mark Stockley blend news, commentary and light-hearted banter while launching a new merch store. The discussion covers real-world harms from AI-generated hoaxes that sent Manila firefighters to a non-existent fire, Albania appointing an AI-made minister, and reports of the so-called 'godfather of AI' being spurned by ChatGPT. They also explore wearable telepathic interfaces like AlterEgo, the rise of AI rights advocacy, and listener support options including ad-free subscriptions and merch purchases.

📘 Google Cloud has published a new ebook, A Practical Guide to Data Science with Google Cloud, aimed at practitioners adopting an AI-first approach across BigQuery, Vertex AI, and Serverless for Apache Spark. The guide emphasizes unified, streamlined workflows enabled by a central notebook experience that blends SQL, Python, and Spark and includes assistive features in Colab Enterprise to generate multi-step plans and code. It explains how a unified data foundation lets teams manage structured and unstructured data together and use familiar SQL to process documents and images. The ebook also offers real-world use cases with linked notebooks so practitioners can run the examples and accelerate delivery.

🚀 Today Google announced 14 startups selected for the Google for Startups Accelerator: AI First program serving the Middle East, North Africa, and Turkey. The cohort addresses challenges across finance, real estate, healthcare, industrial safety, TradeTech, and education, and will receive targeted mentorship, technical training, and product and business support. Participants include Abwab.ai, COGNNA, Distichain, xBites, and Navatech, and the program emphasizes responsible AI to accelerate regional scaling and commercialization.

📘 Bruce Schneier announces his new book, Rewiring Democracy: How AI Will Transform our Politics, Government, and Citizenship, coauthored with Nathan Sanders and published by MIT Press on October 21; signed copies will be available directly from the author after publication. The book surveys AI’s impact across politics, legislating, administration, the judiciary, and citizenship, including AI-driven propaganda and artificial conversation, focusing on uses within functioning democracies. Schneier adopts a cautiously optimistic stance, stresses the importance of imagining second-order effects, and argues for the creation of public AI to better serve democratic ends.

🧠 In episode 66 of The AI Fix, hosts Graham Cluley and Mark Stockley walk listeners through a rapid-fire roundup of recent AI developments, from a ChatGPT prompt that produced an inaccurate anatomy diagram to a controversial Stanford sushi hackathon. They cover a Google Gemini bug that generated self-deprecating responses, criticisms that gave DeepSeek poor marks on existential-risk mitigation, and a debunked pregnancy-robot story. The episode also celebrates a genuine scientific advance: a team of AI agents that designed novel COVID-19 nanobodies, and considers how unusual collaborations and growing safety work could change the broader AI risk landscape.

🔍 The NCSC and the AI Security Institute have broadly welcomed public, bug-bounty style disclosure programs to help identify and remediate AI safeguard bypass threats. They said initiatives from vendors such as OpenAI and Anthropic could mirror traditional vulnerability disclosure to encourage responsible reporting and cross-industry collaboration. The agencies cautioned that programs require clear scope, strong foundational security, prior internal reviews and sufficient triage resources, and that disclosure alone will not guarantee model safety.

⚠️ In episode 65 of The AI Fix, Graham Cluley warns that Microsoft Excel’s new COPILOT function can produce unpredictable, non-reproducible formula results and should not be used for important numeric work. The hosts also discuss a research experiment that created a 500‑AI social network and the arXiv paper Can We Fix Social Media?. The episode blends technical analysis with lighter AI culture stories and offers subscription and support notes.