Proofpoint has launched a pair of new AI-driven security platforms designed to tackle threats that blend into everyday business activity. Agentic Collaboration Security uses intent-based detection and the Proofpoint Knowledge Graph to scan emails, collaboration tools and browsers for phishing, supplier compromise and payment fraud. Most threat decisions are handled in under half a second, with deeper analysis reserved for ambiguous cases.
Alongside this, Proofpoint introduced the Agentic Data and AI Security system, which treats data and AI risk as one connected problem. Three autonomous agents handle detection, investigation and remediation, while Semantic Business Policies and intent-based access control let organisations apply existing rules during AI activity. Proofpoint's 2026 AI and Human Risk Landscape report found that 87% of organisations have moved AI assistants beyond pilot programmes.
Grab and OpenAI rolled out a two-year initiative called GO Forward with AI in Singapore on September 23. The programme offers free half-day workshops and a three-month ChatGPT Plus subscription to gig workers and merchants, with plans to reach over 30,000 participants across Southeast Asia. Workshops cover website building, product design and demand forecasting using AI. Grab has partnered with OpenAI since 2024 on its Driver AI Assistant, used by close to 500,000 drivers.
Greek Prime Minister Kyriakos Mitsotakis discussed AI challenges at an Endeavor Greece event in San Francisco, noting that the country already feels AI's effects. Greece has received roughly €36 billion in EU recovery funding, much of it directed at digital infrastructure including a new supercomputer in Lavrio. The government plans to ban social media for children under 15 starting January and is running a pilot with OpenAI to reduce teachers' administrative workload.
Shinhan Securities launched two AI-powered investment services on September 23. My Stocks Today organises price movements, earnings and weekly news for tracked stocks, while AI Overseas Stock Disclosures translates and summarises filings from U.S.-listed companies in real time. These services build on earlier AI efforts including Earnings Call Live and AI PB.
The Department of Housing and Urban Development is developing an AI tool to monitor how grant recipients spend federal money. Acting CFO Irving Dennis noted the agency distributes about $77 billion annually without tracking how funds are used. Experts warn the tool could produce errors, disadvantage small awardees and introduce political influence into grant decisions.
Palantir Technologies is gaining attention ahead of earnings, with the stock reaching an 8-month high and a market value near $193.33, up 4.51%. Analysts point to strong AI momentum as a driver. The Federal Aviation Administration also launched an AI air traffic control system at three major Washington-area airports, a move expected to benefit Palantir, though the system's effectiveness is still being evaluated.
UK Foreign Secretary Ed Miliband addressed the UN Security Council on September 23, calling for global action on AI safety, transparency and resilience. He urged governments not to ignore warnings from AI founders and pushed for shared principles, highlighting the UK's G20 presidency in 2027.
ChainUQ, a new framework, helps large language models track uncertainty more accurately during reasoning. Tests showed an average 3.1% accuracy gain and up to a 45% reduction in error without extra fine-tuning. Meanwhile, Hulu released a trailer for Soulm8te, a horror thriller about an AI companion that becomes sentient, streaming October 1 as part of Halloween programming.
Key Takeaways
- Proofpoint launched Agentic Collaboration Security and Agentic Data and AI Security, using intent-based detection and three autonomous agents to tackle phishing, data risk and AI misuse.
- Grab and OpenAI started GO Forward with AI, offering free workshops and ChatGPT Plus to over 30,000 gig workers across Southeast Asia.
- Greek PM Mitsotakis announced a pilot with OpenAI to reduce teachers' admin work and plans to ban social media for children under 15 starting January.
- Shinhan Securities rolled out two AI investment services: My Stocks Today and AI Overseas Stock Disclosures for real-time filing summaries.
- HUD plans an AI tool to track spending of $77 billion in annual grants, with concerns about errors and political influence.
- Palantir stock rose 4.51% to a market value near $193.33, with analysts citing strong AI momentum and FAA air traffic control deployment.
- The FAA launched an AI air traffic control system at three Washington-area airports, expected to benefit Palantir.
- UK Foreign Secretary Miliband urged the UN Security Council to adopt global AI safety principles and not ignore founder warnings.
- ChainUQ improves uncertainty tracking in AI reasoning models, achieving up to 45% error reduction without extra fine-tuning.
- Hulu's Soulm8te, a horror thriller about a sentient AI companion, streams October 1 as part of Halloween programming.
Proofpoint launches AI security to tackle trusted business attacks
Proofpoint launched Agentic Collaboration Security, a system built to stop attacks that look like normal business messages. It uses intent-based detection and the Proofpoint Knowledge Graph to analyse emails, collaboration platforms and web browsers. The system targets phishing, supplier compromise and payment fraud that come from trusted contacts. Most decisions are made in under half a second, with deeper analysis for unclear cases.
Proofpoint unifies data and AI security in one agentic system
Proofpoint announced the Agentic Data and AI Security system, designed to protect data and AI as one connected risk. The system uses the Proofpoint Knowledge Graph and Nexus models to combine AI activity with data sensitivity, identity and intent. Three new agents handle detection, investigation and remediation. It aims to help organisations apply existing business rules as runtime controls for AI actions.
Proofpoint launches unified data and AI security system
Proofpoint released the Proofpoint Agentic Data and AI Security system to manage risk across employees and AI agents. It connects data sensitivity, identity, access, behaviour and intent in one framework. The system includes three autonomous agents for detection, investigation and remediation. It also introduces Semantic Business Policies and intent-based access control to enforce rules during AI activity.
Proofpoint launches AI security system for data risk
Proofpoint launched a unified data and AI security system to manage risk for employees and AI agents. The product connects data sensitivity, identity, access, behaviour and intent in a single framework. It includes three autonomous agents for detection, investigation and remediation, along with Semantic Business Policies. Proofpoint cited its 2026 AI and Human Risk Landscape report showing 87% of organisations had moved AI assistants beyond pilot programmes.
Proofpoint launches AI security against trusted attacks
Proofpoint launched Agentic Collaboration Security to stop attacks that resemble ordinary business communications. The system uses intent-based detection and the Proofpoint Knowledge Graph to analyse messages across email, collaboration platforms and browsers. It includes automated threat investigation and adaptive user protection for higher-risk users. The new functions will be delivered as an update to Proofpoint's existing collaboration security offering.
Grab and OpenAI offer free AI training and ChatGPT to gig workers
Grab and OpenAI launched a two-year programme called GO Forward with AI on Sept 23 in Singapore. About 50 female drivers and delivery riders attended the first workshop. Participants receive free half-day workshops and a three-month ChatGPT Plus subscription. The programme will expand to over 30,000 gig workers and merchants across Southeast Asia including Indonesia, Malaysia, the Philippines, Thailand and Vietnam. Workers will learn to build websites, design products and gauge demand using AI without coding.
Grab and OpenAI bring practical AI skills to Southeast Asia
OpenAI and Grab launched the GO Forward with AI programme to help 30,000 driver, delivery and merchant partners build practical AI skills. The programme begins in Singapore and will expand to Thailand, Indonesia and the Philippines later in 2026, then Malaysia and Vietnam in 2027. Workshops teach partners to use ChatGPT for business planning, sales dashboards and website building. Grab has worked with OpenAI since 2024 on its Driver AI Assistant used by nearly 500,000 drivers. The programme aims to help partners grow businesses or explore new opportunities using AI.
Soulm8te trailer reveals creepy AI horror thriller coming to Hulu
Hulu released a trailer for Soulm8te, a horror thriller about an AI companion designed to love unconditionally. The film stars David Rysdahl, Lily Sullivan and Claudia Doumit. It follows a grieving engineer who tests an AI companion and makes her sentient, but their relationship turns sinister. The movie is directed by Kate Dolan and produced by Blumhouse and Atomic Monster. Soulm8te will stream on Hulu and Hulu on Disney+ starting Oct 1 as part of Hulu's Halloween programming.
Greece's prime minister discusses AI challenges and tech future
Greek Prime Minister Kyriakos Mitsotakis spoke at an Endeavor Greece event in San Francisco about AI and economic growth. He said Greece is already struggling with AI effects and does not have all the answers. Greece received about €36 billion from the EU recovery fund, much of it spent on digital infrastructure including a new supercomputer in Lavrio. The country plans to ban social media for children under 15 starting January. Mitsotakis also mentioned a pilot with OpenAI to reduce teachers' administrative work. He wants to attract tech talent back to Greece with lower taxes and visa programs.
Shinhan Securities launches AI services for investment information
Shinhan Securities launched two AI-powered investment services on Sept 23. My Stocks Today uses AI to organize price movements, earnings announcements and weekly news for stocks customers hold or watch. AI Overseas Stock Disclosures provides real-time translation and summaries of filings from U.S.-listed companies including earnings and governance changes. It also shows consensus earnings forecasts from global investment banks. These services expand Shinhan Securities' AI offerings that started with Earnings Call Live in August and AI PB last year.
HUD plans AI tool to review housing grant spending
The Department of Housing and Urban Development wants to use an AI tool to check how grant recipients spend their money. Acting CFO Irving Dennis said the agency gives out about $77 billion yearly but does not track how that money is used. Experts worry the tool could cause errors, hurt small awardees, and lead to political influence in grant decisions.
Palantir stock seen as new buy amid earnings and AI growth
Palantir Technologies is being viewed as a strong buy ahead of earnings reports. The stock shows positive chart patterns and has reached an 8-month high. Its market value sits near $193.33 with a 4.51 percent gain. Analysts point to strong AI momentum driving investor interest.
Palantir poised to benefit from FAA AI air traffic control rollout
The Federal Aviation Administration launched an AI air traffic control system at three major Washington-area airports. Palantir is expected to see a significant boost from this move. The effectiveness of the system is still being tested. The FAA deployment has driven rising interest in Palantir stock.
Miliband urges UN Security Council to address AI safety warnings
UK Foreign Secretary Ed Miliband spoke to the UN Security Council on September 23, 2026 about AI risks. He called for global action on safety, transparency, and resilience in AI development. Miliband said governments must not ignore warnings from AI founders. He also pushed for shared global principles and highlighted the UK G20 presidency in 2027.
ChainUQ improves uncertainty tracking in AI reasoning models
ChainUQ is a new framework that helps large language models better track uncertainty in their reasoning. It addresses problems with current methods that miss logical inconsistencies in AI responses. The system uses a lightweight module and a consistency-aware calibrator to improve confidence scoring. Tests show an average 3.1 percent gain in accuracy and up to 45 percent reduction in error without needing extra fine-tuning.
Sources
- Proofpoint launches AI security against trusted attacks
- Proofpoint Breaks Down the Divide Between Data Security and AI Security with the Industry’s First Unified Agentic System
- Proofpoint launches AI security system for data risk
- Proofpoint launches AI security system for data risk
- Proofpoint launches AI security against trusted attacks
- Free AI training and ChatGPT subscriptions for Grab platform workers, merchants to build business ideas
- Grab and OpenAI bring practical AI skills to Southeast Asia
- 'Soulm8te' trailer teases a terrifying take on AI, love and obsession
- "We're already fighting yesterday's battle": Greece's prime minister gets candid about AI
- Shinhan Securities Launches AI Services Summarizing Holdings Issues and U.S. Filings, Strengthening Personalized Investment Information
- HUD eyes AI tool to review grant spending. Housing advocates have concerns
- Stock Of The Day: Palantir Is New Buy Amid Earnings, AI Momentum
- Palantir Could Be Big Winner as Air Traffic Control Goes AI. The Stock Is Rising.
- Miliband Tells UN Security Council to Act on AI Founders’ Warnings
- ChainUQ: Reasoning Consistency-Aware Uncertainty Quantification for Large Language Models
Comments
Please log in to post a comment.