Brett Hurt publishes Aspen debate on who sets AI's values

Why it matters

Hurt's shift after the data.world sale shows how the AI governance debate is moving closer to operators who built the enterprise data systems these models depend on.

Brett Hurt (@brettahurt) released a recording of his Aspen Institute discussion with futurist Jamie Metzl on August 6th, extending the data.world co-founder's post-exit shift from building enterprise software to arguing about the values that should govern increasingly capable artificial intelligence.

The conversation was recorded on July 24th at the Aspen Institute, where Hurt and Metzl appeared with moderator Chadia El Meouchi Naoum. The event centered on a familiar list of AI risks, including job displacement, privacy, security and misinformation, while also considering the technology's potential role in scientific discovery, healthcare and environmental protection.

Hurt's framing carries more weight than a generic call for responsible AI. He has spent his career building the data systems that now sit beneath corporate AI deployments. Hurt founded Coremetrics, later acquired by IBM, and took Bazaarvoice public before co-founding data catalog and governance provider data.world. ServiceNow completed its acquisition of data.world in July 2025 and said it planned to use the technology to give AI agents better context about enterprise data. Financial terms were not disclosed.

That history creates the central tension in Hurt's new work. Enterprise AI depends on clean data, defined ownership and enforceable access rules. Moral judgment is harder to encode. The episode description says the speakers considered why technologies without values require a human moral compass, and how advances across AI and other fields could produce either widespread abundance or severe harm.

Hurt's post-exit platform

Hurt now describes Love Conquers Fear as a holding company focused on technology, consciousness and what he calls an "Age of Abundance for All." The organization includes a book and podcast built around the premise that AI, robotics, quantum computing and brain-computer interfaces are advancing faster than the social institutions expected to manage them.

The Aspen recording gives that project a founder-led thesis: technical capability alone cannot decide which outcomes companies should pursue. That puts responsibility back on executives, investors and engineers choosing what gets funded, optimized and deployed.

The argument also reflects Hurt's move from operator to public advocate after the data.world sale. His earlier companies were designed to collect, organize and apply information at commercial scale. His current project asks what happens when those systems become capable enough to influence work, institutions and individual decisions without carrying an independent moral framework.

Metzl's experiment with GPT-5

Jamie Metzl (@JamieMetzl) approached the discussion from a different direction. His 2026 book, The AI Ten Commandments, is presented as a collaboration with GPT-5 that draws principles from religious, philosophical, Indigenous and humanist traditions.

The authorship label is deliberately provocative, though Metzl's own account places control with the human author. In an official question-and-answer page, the book describes GPT-5 as a tool for identifying connections, testing language and generating alternatives, while Metzl retained responsibility for the thesis, values and final decisions.

That distinction matters for companies adopting similar systems. A model can compare policies, surface precedents and generate options at a scale no individual employee can match. It cannot accept legal liability, answer to a board or live with the consequences of a decision. The people deploying the model retain those obligations, even when an automated output appears confident or comprehensive.

Hurt and Metzl arrive at the same operational problem from different careers. Hurt built companies around making data useful inside organizations. Metzl is testing whether a model can help synthesize humanity's existing ethical traditions. Both approaches still require a person to choose the objective, reject bad outputs and remain accountable for the result.

The August 6th release publishes a July 24th event rather than a new product or policy proposal. Its significance lies in Hurt's post-acquisition direction. After helping build the data layer for enterprise AI, he is using his next platform to argue that the harder constraint will be the judgment of the people controlling it.