Former Intel CEO Pat Gelsinger, who has retired from a 40-plus-year career in semiconductors, has unveiled a new initiative — the Flourishing AI (FAI) benchmark — designed to assess how well large language models align with fundamental human values, according to an article in The New Stack. The project is being implemented in partnership with technology company Gloo, which Gelsinger joined as an investor 10 years ago.
The FAI is based on the Global Flourishing Study, a study conducted by Harvard University and Baylor University to measure the well-being of people in different countries. The test includes seven categories: "Character and Integrity", "Close Social Connections", "Happiness and Life Satisfaction", "Meaning and Purpose", "Mental and Physical Health", "Financial Stability", and "Faith and Spirituality". These parameters are used to test whether the behavior of AI is in line with universal human guidelines.
The FAI initiative is in its early stages, and the development team has not yet interacted directly with AI model providers. However, Pat Gelsinger, the project’s founder, expressed hope that the new benchmark will open the way for a broader discussion in the industry. According to him, if AI models improve on the seven dimensions of well-being, it will be a significant achievement for all of humanity.