The Vatican is actively considering the ethical and societal implications of AI in a rapidly changing world, emphasizing the importance of trust as a foundational value. The discussion highlights challenges such as declining trust in institutions and fragmented media, advocating for transparency, consistent communication, and empowering credible voices to build and maintain trust in AI technologies.

AI-powered app Auni, developed by Fastagger and running on Microsoft Azure, is helping Kenyan small business owners like Mandevu Beard Care analyze mobile money transactions to gain business insights. Integrated with Safaricom’s dominant M-Pesa mobile payment system, Auni leverages AI tools including OCR and small language models to simplify financial management and decision-making for entrepreneurs in Kenya’s growing middle class. This innovation exemplifies how AI is driving digital transformation and financial inclusion in Kenya’s vibrant mobile money ecosystem.

Agoda has open-sourced APIAgent, a tool that converts any REST or GraphQL API into a Model Context Protocol (MCP) server without requiring code. This innovation aims to simplify AI agents' access to data by enabling seamless communication with existing APIs.

The article presents a tutorial on building a human-in-the-loop AI travel booking agent using LangGraph and Streamlit, emphasizing explicit user approval before execution. The agent generates a structured travel plan and pauses to seek user confirmation, treating the user as an active collaborator rather than a passive observer. This approach enhances transparency and control in AI decision-making processes.

Alibaba's Qwen team has released Qwen3.5-397B-A17B, a sparse Mixture-of-Experts large language model with 17 billion active parameters and a 1 million token context window, enhancing both reasoning power and efficiency. This native vision-language model is designed to advance AI agent capabilities by integrating massive scale with multimodal understanding.