Former Intel CEO Ji Zinger is training a "Christian-oriented large language model": hoping to accelerate the return of Jesus
Former Intel CEO Patrick Gelsinger moved to Gloo to promote Christian AI and "Human Prosperity Benchmarks," raising new issues on AI ethics and industrial layout.
(Preliminary summary: Intel’s radical proposal: "Split off the market" foundry and design departments, can Trump and Nvidia join forces to reshape American chip hegemony? )
(Background supplement: Nvidia announced a US$5 billion investment in Intel: joint cooperation in chip design, Intel rose 30% before the market opened )
Contents of this article
Patrick Gelsinger, the former CEO of the US chip maker Intel (Intel), revealed in the middle of this year that leaving Intel was not his intention, but a decision made by a third party. He deeply regrets that he cannot complete the "IDM 2.0" wafer foundry strategy that is being promoted.
But now he seems to have found a new direction. According to "Fortune", Gelsinger is currently the CEO of Gloo, focusing on building "Christian-oriented large-scale language models", hoping to use AI to accelerate the "return of Christ" and spark discussions.
From chip giant to Holy Light AI
Gelsinger has always considered himself a devout Christian. In the past, he published "Balancing Your Family, Faith & Work" and regarded Silicon Valley as a missionary field.
After leaving Intel, he chose to join Gloo, which calls itself a "connected faith ecosystem." According to the official website, the platform serves more than 140,000 religious and non-profit leaders. Its core products "Christian Alignment Large Language Model (CALLM)" and "Kingdom Alignment Large Language Model (KALLM)" allow churches to use sermon content to train AI assistants to handle pastoral care, discipleship training and administrative management.
Gelsinger called AI the "Gutenberg moment" after printing at a seminar at Colorado Christian University, and asked:
"How can we shape AI as a technology so that it can truly become a powerful embodiment and expression of the church?"
He hopes the answer lies in writing biblical ethics into the weight of the model, "so that every inference responds to human dignity."
"Values Alignment" and Human Prosperity Benchmark
In order to quantify the standard of "goodness", Gloo refers to the concept of Harvard's "Human Prosperity Project" and launches the Human Prosperity AI Benchmark, which evaluates model contributions based on mental, emotional and other indicators.
Internal testing shows that although mainstream models perform well in translation and summarization, they struggle to provide in-depth faith dialogue. Therefore, Gloo held an AI hackathon with a total prize of more than 250,000 US dollars, calling on developers to create "values alignment" tools, emphasizing that funds will be concentrated on projects that can measure human well-being.
Unknown variables in industry and social impact
In the short term, the addition of Gelsinger will give Gloo a strong boost in fundraising and brand recognition, and will also open up a new battlefield for the "values competition" for the AI industry. However, in the long run, once religiously oriented models are integrated into public decision-making or commercial services on a large scale, how to balance religious freedom and the needs of diverse societies will inevitably test the regulatory framework.
At a time when unified ethical standards for AI have yet to be established, Gloo’s experiment provides a window into the integration of faith and technology, and also exposes people’s differences on the concept of “neutral technology.” When theological terms such as "Christ's Second Coming" and "Antichrist" officially enter the product roadmap, Silicon Valley may be heading towards an unprecedented intersection: it may not only bring new tools for human welfare, but may also push different beliefs into sharper tension.
Whether this experiment can verify a new formula for "human prosperity" or just add religious vocabulary to a technological utopia remains to be determined by time.
