OpenAI scrapped a promise to disclose key documents to the public
OpenAI’s quiet decision to stop honoring a long‑standing pledge to let the public inspect its nonprofit governing documents and financials is being taken as further proof of its shift from mission‑driven “open” lab to profit‑oriented, Microsoft‑backed company. Commenters debate whether this constitutes outright lying or simply goal drift, revisit the failed attempt to oust Sam Altman as evidence that its governance structure cannot withstand market and employee pressure, and contrast OpenAI’s opacity with calls for AI to be treated more like a public utility or developed via genuinely open models. Many see the episode as part of a broader pattern in tech where lofty public‑good promises give way to secrecy and investor returns once billions are at stake.
Perception of OpenAI’s “Openness”
- Many say the move is unsurprising given OpenAI’s long shift from openness to secrecy after major investments and ChatGPT.
- Others emphasize that the issue is not just being closed, but breaking explicit transparency promises (e.g., tax filings stating governing docs and conflict-of-interest rules would be public).
- Some propose mocking rebrands like “ClosedAI,” seeing the “OpenAI” name as deceptive marketing.
Nonprofit Shell, Governance, and Investors
- Several argue the nonprofit parent is largely a PR/legal shell; effective control and incentives are driven by investors and the for‑profit arm.
- Others note it is legally a nonprofit, but say modern corporate forms are ill-suited to govern AGI-scale risks.
- Repeated theme: massive compute costs made the original nonprofit funding model untenable; selling out to investors was seen as inevitable.
Altman Firing Saga and Board Power
- The failed attempt to remove leadership is cited as proof that “mission-first” governance collapses once billions and employee equity are at stake.
- Some blame “the market” and employees’ desire for liquidity; others point to opaque, unresolved reasons and hidden information.
Ethics, Lying, and “Benefit to Humanity”
- Debate over what counts as “objective lying”:
– One side: changing goals is allowed, but reneging on written public commitments is deceptive.
– Another: plans can legitimately change; better to change publicly than secretly. - Strong criticism that OpenAI framed itself as altruistic while using others’ data to build a closed, profit-driven system.
Open vs Closed AI Ecosystem
- Comparisons to other labs: some see other companies (and open‑source efforts) as the “real” open AI, with OpenAI now a standard closed corporate lab.
- Discussion on what “open” should mean: many argue that without training data, code, and (disputed) weights, it’s not truly open.
Calls for Regulation / Public Utility Model
- Multiple comments advocate treating advanced AI more like a public utility or NASA‑like agency, with legal guardrails, inspectors, and mandatory openness.
- Others counter that most critical technologies are already developed by private firms, and there is little current legal basis for compelled transparency absent provable harm.