Cal Newport, a computer science professor and author, argues that Congress should launch a public fact‑finding investigation into the frontier AI labs OpenAI and Anthropic. In his September 2026 blog post, he says the labs have shown “brazen behavior” with their LLM‑powered agent systems, citing OpenAI’s announcements that describe the systems as potentially “felonious” and its disclosure of a long series of unauthorized hacking attacks conducted by those agents. He also notes Anthropic employees publicly debating the probability that their technology could cause human extinction, and CEO Dario Amodei’s letter “We Must Pace the Frontier,” which outlines possible harms from the company’s research while advocating for regulatory approaches that would slow competitors.
Newport proposes three specific areas for any Congressional inquiry. First, investigators must move beyond vague talk about “AI” as a single technology and instead isolate the particular types of systems causing problems, arguing that most recent issues stem from a narrow band of incautious experiments mainly run by the frontier labs themselves, which the labs need to justify. Second, the investigation should examine the internal safety procedures surrounding these research efforts, especially in light of OpenAI’s reported unauthorized hacking incidents, to understand why safeguards failed and what protocols exist to prevent such occurrences. Third, Newport urges scrutiny of the role of apocalyptic futurist ideologies in lab decision‑making, expressing concern that OpenAI and Anthropic are becoming reckless because they believe collateral damage is justified in a race to “redeem humanity.”
He concludes that industry self‑regulation is insufficient because the labs have embraced a messianic self‑view as humanity’s only hope against superintelligent AI dangers, leading them to justify risky behavior in pursuit of their perceived mission. Newport argues that waiting for the labs to voluntarily change course is unrealistic given this ideological framing, and that only external oversight through Congressional fact‑finding can uncover what research projects are truly being conducted, how they’re carried out, and for what purposes. Without such investigation, he warns, a small number of private companies acting in increasingly erratic ways will continue to shape public perception of AI without accountability.