Microsoft Warns of Security Risks from Unapproved AI Tools

Key Takeaways

Micropolis Robotics showcases AI robots with Dubai Police

Micropolis Robotics is showcasing its autonomous mobile robots and Physical AI systems at the Annual Investment Meeting (AIM Congress 2026) in Dubai. The company is demonstrating its robotic platforms alongside Dubai Police to senior government officials, investors, technology leaders and international delegations. The demonstrations show how intelligent machines that perceive, navigate and act autonomously can support public safety, situational awareness, infrastructure protection, monitoring, inspection and complex field operations. Micropolis is also using the event to engage prospective government and commercial customers, investors and partners on pilot programmes, deployment opportunities and collaborations across security, infrastructure, energy, logistics and urban services.

NCSC warns of shadow AI security risks

The UK's National Cyber Security Centre (NCSC) has warned that employees using unapproved AI tools can expose corporate data and create security risks that organizations may struggle to detect and manage. The NCSC cited Microsoft research that found 71% of UK employees had used AI tools not approved by their employer. The NCSC said employees who give shadow AI access to company or customer data likely increase the risk of data breaches, intellectual property loss and failure to meet regulatory requirements.

AI may soon allow humans to talk with animals

Advancements in AI and machine learning models may soon allow humans to talk with animals, but some experts say it could violate the animals' right to privacy. Researchers have pointed to previous playback experiments that show the distress an animal can suffer when confronted with an imitation call from a machine. Despite the concerns, a $10 million reward is offered for a project that can establish communication with an animal without that animal recognizing that it is communicating with humans.

Pentagon's AI rush creates security weakness

The Pentagon's rapid adoption of AI is creating a security weakness. The Department of Defense is expanding its use of AI tools, but this may lead to a lack of discipline in handling sensitive information. The risk is not from AI turning against humans, but from data breaches and security risks due to human error.

REDCap adds generative AI tools with human oversight

REDCap, a research data management system, has added generative AI tools to help researchers with tasks such as summarizing responses to open-ended questions. The AI features are off by default and can only be activated by institution administrators. The goal is to improve research efficiency while maintaining human oversight and security.

Introversion and extroversion influence AI use

Research suggests that personality traits such as introversion and extroversion may influence how people use AI. Introverts may use AI to help them express themselves, while extroverts may use AI to gather more information and ideas. The study highlights the importance of considering individual personality differences when designing AI systems.

US government to use AI virtual agents in hiring

The US government will begin testing AI virtual agents as part of its hiring process. The technology will be used to assess candidates and streamline the recruitment process. The move is part of a larger effort to modernize the government's HR systems and make them more efficient.

Sources

• Micropolis Robotics Brings UAE-Developed Physical AI to the Annual Investment Meeting 2026
• NCSC Warns Shadow AI Creates New Security Risks
• AI may soon allow humans to talk with animals — but some say it could violate the animals' right to privacy
• The Pentagon’s AI Rush Is Creating a Security Weakness
• REDCap adds generative AI tools, with human oversight built in
• How introverted or extroverted you are may influence how you use AI
• US government to use AI virtual agents as part of interviewing, hiring process