This weekly update on accessibility and inclusion in agentic AI reveals limited but promising developments. The most significant advancement involves on-device voice AI technology that enables secure, real-time transcription and translation capabilities. This breakthrough supports multilingual accessibility, making AI agents more available to people worldwide regardless of language barriers. The technology includes built-in features for handling specialized terminology and noisy environments, which benefits users with hearing differences and professionals in challenging work settings. Additionally, organizations are increasingly recognizing the need for human-in-the-loop oversight of AI agents, with 68% rating this supervision as essential or very important. This approach helps ensure agents work fairly and safely for all users. However, the search results indicate that most organizations currently lack formal strategies for secure agent deployment, affecting how accessible and inclusive these systems can become. The focus remains on building foundational governance structures that will eventually support more equitable AI agent access across different communities and industries.