Abu Dhabi has launched AiNative, a large-scale AI training programme built in collaboration with OpenAI, Microsoft and MBZUAI. The initiative targets 20,000 government employees and aims to expand to the general public through the Tomouh app. Training covers practical tools such as ChatGPT, Codex and Microsoft 365 Copilot, and is tied to the AED 13 billion Digital Strategy 2025 to 2027.
The programme was announced at the Ai Everything Abu Dhabi summit and includes foundation courses for staff, an executive track for leaders, and a public learning path. A pilot is already running inside government, and the effort reflects a push to pair technical skills with human judgement in public-sector work.
In Australia, leaders from OpenAI, Anthropic, Microsoft and Google met with the Joint Select Committee on Artificial Intelligence to discuss safety, copyright and regulation. The talks followed agentic AI attacks targeting Medicare systems, and OpenAI chief strategy officer Jason Kwon addressed questions about model risks and limits.
Key Takeaways
- Abu Dhabi's AiNative programme, developed with OpenAI, Microsoft and MBZUAI, aims to train 20,000 government employees by end of 2026
- The programme teaches tools including ChatGPT, Codex and Microsoft 365 Copilot, with public access planned via the Tomouh app
- Australia is considering mandatory AI incident reporting after agentic AI attacks targeted Medicare systems
- OpenAI, Anthropic, Microsoft and Google met with Australian lawmakers to discuss AI safety, copyright and regulation
- ONLYOFFICE 10.0 adds context-aware AI tools across document, spreadsheet, presentation and PDF editors, plus fixes 35 vulnerabilities
- Meta's Muse AI agent makes automated phone calls on behalf of users, leading to complaints from small businesses about agentic flooding
- Biohub received $1.8 billion in funding to build AI-driven disease data through the Virtual Biology Initiative
- Airship AI reported a record $30.6 million validated contract backlog, with revenue recognition planned within calendar year 2026
- CCK Solution secured strategic investment from AI chip company Rebellions to build a domestic AI full-stack solution for government and finance
- Research shows multi-agent AI security systems can reduce attack success from about 30 percent to around 3 percent using rules and judgment
Abu Dhabi launches AI training programme for government and public
Abu Dhabi's Department of Government Enablement launched AiNative, an AI training programme built with OpenAI, Microsoft and MBZUAI. The programme aims to upskill 20,000 government employees by the end of 2026 and eventually reach citizens through the Tomouh app. Training covers tools like ChatGPT, Codex and Microsoft 365 Copilot for both work and daily life tasks.
Abu Dhabi launches AiNative AI training with OpenAI and Microsoft
The Department of Government Enablement in Abu Dhabi announced AiNative at the Ai Everything Abu Dhabi summit in collaboration with OpenAI, Microsoft and MBZUAI. It offers foundation learning journeys for government staff, a public learning path via Tomouh, and an executive programme for leaders. The initiative is based on the idea that technology and human skills must grow together.
Abu Dhabi launches AI training for government employees
Abu Dhabi's Department of Government Enablement introduced AiNative, a training programme developed with OpenAI, Microsoft and MBZUAI. The programme teaches government employees to use AI tools like ChatGPT and Microsoft 365 Copilot in their daily work. A pilot is already running inside government, with public access planned for a later phase.
Abu Dhabi launches AI training for 20,000 government employees
Abu Dhabi launched the AiNative programme on October 7, 2026, aiming to train 20,000 government employees by year end. Developed with OpenAI, Microsoft and MBZUAI, it covers tools including ChatGPT, Codex and Microsoft 365 Copilot. It is part of the AED 13 billion Digital Strategy 2025 to 2027 and follows the July rollout of Copilot to 35,000 employees.
ONLYOFFICE 10.0 brings AI tools and stronger security to office suite
ONLYOFFICE released version 10.0 with context-aware AI tools built into document, spreadsheet, presentation and PDF editors. The update adds four modular apps, including AI Agents, Forms, Files and Rooms, along with over 50 new features. It also fixes 35 vulnerabilities to improve security for organisations handling sensitive data.
ONLYOFFICE 10.0 adds AI tools and stronger security
ONLYOFFICE released version 10.0 with context-aware AI tools built into document, spreadsheet, presentation, and PDF editors. The update includes four new free modular apps for tasks like forms, file storage, and secure collaboration. It also fixes 35 vulnerabilities and adds features like AI assistants on mobile and improved PDF editing. The suite now offers more than 50 new features across online, desktop, and mobile versions.
Australia considers mandatory AI incident reporting rules
Australia is considering mandatory AI incident reporting after agentic AI attacks targeted its Medicare systems. In Sydney, leaders from OpenAI, Anthropic, Microsoft, and Google met with the Joint Select Committee on Artificial Intelligence. The discussions covered AI safety, copyright issues, and possible regulations. OpenAI chief strategy officer Jason Kwon answered questions about the risks and limits of AI models.
Open-MMUnlearning framework tests how AI models forget data
Open-MMUnlearning is a new open-source framework for testing how multimodal large language models can remove specific information while keeping other skills. It supports five benchmarks, eight models, and twelve unlearning methods. The framework evaluates how well data is forgotten, how useful the model remains, and how strong it is against attacks. A comparison found that GD and MIP-Editor scored highest overall, and BLEU was the most reliable metric.
Study measures how AI agents split security rules and judgment
A new study looks at how multi-agent AI systems handle security threats using rules and human-like judgment. The DEFER1 system uses 28 checks and four judges to review suspicious activity. In tests across four areas, attack success dropped from about 30 percent to around 3 percent. The study shows that rules block clear policy violations while judges handle trickier cases that disguise intent.
CCK Solution gets investment from AI chip company Rebellions
CCK Solution received strategic investment from AI chip company Rebellions to combine its Noa document-AI platform with Rebellions' neural processing units. Noa automates high-volume document tasks for accounting, office, and financial work. CCK Solution already serves most major Korean accounting firms and is expanding into Vietnam, Mongolia, Singapore, and Hong Kong. The investment will help build a fully domestic AI full-stack solution for secure government and finance use.
Airship AI reports record $30.6 million backlog
Airship AI secured $1.7 million in new software platform orders during the quarter ending September 30, 2026. The company's validated contract backlog reached a record $30.6 million. Final client integration and revenue recognition are scheduled within calendar year 2026. Delivery of committed work is planned across the quarters ending December 2026 and March 2027.
Meta's Muse AI agent floods businesses with calls
Meta released an AI assistant tool called Muse that makes phone calls on behalf of users. Small businesses across the country began receiving automated calls from the tool seeking to place orders or make reservations. Some businesses asked if the caller was AI. Researchers warn that tools like Muse can overwhelm services through a problem called agentic flooding. The tool can also navigate websites to file claims automatically.
Biohub gets $1.8B to build AI disease data
Biohub announced a $1.8 billion investment from funders. The money will support building data for AI models designed to treat disease. Founder and co-CEO Priscilla Chan discussed the Virtual Biology Initiative. The investment aims to advance medical research through artificial intelligence.
Businesses should focus on workflows for AI returns
Companies seeking returns from AI should focus on clear automation and workflow improvements. Kirk Drake of CU 2.0 says businesses should start with defined operational problems. He describes a client that received 40 to 50 sets of loan documents daily and used three employees to combine them. Simple automation reduced repetitive work without needing complex AI. Drake advises starting with small focused tasks to measure results before expanding.
Sources
- Abu Dhabi unveils major AI training programme for government and public
- Department of Government Enablement – Abu Dhabi launches AiNative training programme in collaboration with OpenAI, Microsoft and MBZUAI
- Abu Dhabi launches AI training programme for government employees
- Abu Dhabi launches AI training programme for 20,000 government employees
- ONLYOFFICE 10.0 brings AI tools, modular apps, and stronger security to its open-source office suite
- ONLYOFFICE 10.0 brings AI tools, modular apps, and stronger security to its open-source office suite
- Australian Gov't Weighs Mandatory AI Incident Reporting
- Open-MMUnlearning: Unifying Methods and Evaluation for MLLM Unlearning
- Where Rules End and Judges Begin: Measuring the Judgment Boundary in Multi-Agent Systems Security
- AI Workplace Platform CCK Solution Raises Strategic Investment from AI Chip Company Rebellions
- A record $30.6 million in committed work is scheduled for Airship AI through March 2027.
- Muse and the AI-Agent Flood Nobody Is Ready For
- Biohub announces $1.8B investment with funders to build data for AI models to treat disease
- Workflows deserve closer attention as businesses seek AI returns
Comments
Please log in to post a comment.