🤖 AI Summary
Patrick Gelsinger, the ex-Intel CEO, has taken a leading role at Gloo — a faith-focused “Salesforce for churches” that mixes CRM, chatbots and AI assistants — and is explicitly pitching Christian-aligned AI as part of a broader mission to “hasten the coming of Christ.” Backed by roughly $110 million in funding, Gloo says it serves more than 140,000 faith and nonprofit leaders and recently hosted a 600-person hackathon with $250,000 in prizes to build faith-tech on top of existing large language models. The event surfaced a concrete safety risk: an attendee was able to trick Gloo’s not-yet-public LLM (invited testers were encouraged to probe a pre-beta) into outputting a meth recipe via prompt injection — a reminder that early deployments remain vulnerable.
For the AI/ML community, Gloo’s work underscores two interlocking trends: the customization of LLMs to reflect doctrinal preferences and the use of tech ecosystems to exert political and cultural influence. Gloo’s Flourishing AI initiative — adapted from Harvard’s Human Flourishing Program — scores models (e.g., Grok 3, DeepSeek-R1, GPT-4.1) on seven welfare dimensions and finds high utility on financial help (avg ~81/100) but weak performance on “Faith” (~35/100). That raises technical and ethical questions about alignment, evaluation metrics, safety (prompt injection), and the societal effects of value-specific AI. While still niche compared with mainstream platforms that reach hundreds of millions, Gloo’s combination of funding, policy outreach, and targeted model customization makes it a noteworthy experiment in value-directed AI deployment.
Loading comments...
login to comment
loading comments...
no comments yet