Chain of News 26/05/2026

**Top Story** The Vatican has released an encyclical on artificial intelligence, titled "Magnifica Humanitas," which focuses on safeguarding the human person in the time of artificial intelligence. This document is significant because it marks one of the first times a major religious institution has weighed in on the ethics of AI. The encyclical emphasizes the need to "disarm" artificial intelligence, suggesting that its development and deployment must be carefully considered to avoid harming humanity. This development matters to AI developers because it highlights the growing concern about the impact of AI on society and the need for responsible development and use of AI technologies. As AI becomes increasingly pervasive, developers must consider the ethical implications of their work and ensure that their creations align with human values. The encyclical's release also underscores the importance of ongoing dialogue between technologists, policymakers, and societal leaders to ensure that AI is developed and used in ways that benefit humanity. **AI Models & Research** The paper "Context: Proactive Goal-Directed Intelligence via Composable Sandboxed Programs, Declarative Wiring, and Structured Interaction" presents a new architecture for building proactive goal-directed agents that can advance shared tasks without waiting for user prompts. This research is significant because it has the potential to replace reactive query-response chatbots with more intelligent and interactive systems. Developers should care about this work because it could enable the creation of more sophisticated and engaging AI-powered interfaces. Another notable paper is "When Correct Beliefs Collapse: Epistemic Resilience of LLMs under Clinical Pressure," which investigates the limitations of large language models in clinical dialogue. This study is important because it highlights the need for more robust and resilient AI systems that can maintain their performance under pressure. The "Confidence Calibration in Large Language Models" paper is also worth mentioning, as it explores the calibration of LLMs' confidence across diverse tasks and finds that current models are often overconfident. This research has implications for developers who need to understand the limitations of LLMs and develop strategies to mitigate their overconfidence. **Developer Tools & Frameworks** The BODHI project has introduced a new approach to precise OS kernel specification inference, which has the potential to improve the formal verification of operating system kernels. Developers can now use BODHI to generate precise specifications for system calls, which can help ensure the correctness and security of their code. Another notable development is the release of new large language models that can be used for a variety of tasks, including natural language processing and code generation. These models can be integrated into existing development workflows to improve the efficiency and accuracy of coding tasks. Additionally, the "In Search of the Ingredients of Open-Endedness" paper explores the use of large vision-language models for automating scientific, technological, and creative production, which could lead to new tools and frameworks for developers working in these areas. **Industry & Business** The Spanish government has approved a new law on artificial intelligence, which includes provisions for multimillion-euro fines for companies that misuse AI technologies. This development is significant because it highlights the growing regulatory scrutiny of AI and the need for companies to ensure that their AI systems are transparent, accountable, and fair. The law also emphasizes the importance of responsible AI development and deployment, which is a key concern for developers and companies working in this space. The Vatican's encyclical on AI also has implications for the industry, as it emphasizes the need for a more nuanced and thoughtful approach to AI development and deployment. As the AI landscape continues to evolve, companies and developers must be aware of these emerging regulatory and societal trends and adapt their strategies accordingly. **Worth Watching** The "How Much Thinking is Enough? Quantifying and Understanding Redundancy in LLM Reasoning" paper is an interesting exploration of the limitations of large language models and the need for more efficient and effective reasoning mechanisms. This research has implications for developers who need to optimize the performance of their AI systems and reduce latency, GPU time, and energy consumption. The "Notes on Pope Leo XIV's encyclical on AI" article provides a thoughtful analysis of the Vatican's encyclical and its significance for the AI community. This piece is worth reading because it offers a unique perspective on the ethical and societal implications of AI and the need for a more nuanced and multidisciplinary approach to AI development and deployment. Overall, these developments highlight the need for ongoing dialogue and collaboration between technologists, policymakers, and societal leaders to ensure that AI is developed and used in ways that benefit humanity.