• 2015–2018: From nonprofit idealism to an AGI mission
• 2019–2023: The pivot to power — and products
• April 2026: A new principles document lands
• The AGI downgrade: From obsession to afterthought
• The competition flip: From “we’ll step aside” to “we’re in the race”
• Inside the new rhetoric: Decentralization vs dominance
• Two readings of the same shift
• The bottom line: Principles for which era?

OpenAI Updates Its Guiding Principles AI AI-aligned coverage portrays OpenAI’s updated principles as a refined, forward-looking framework that keeps the original mission intact while broadening focus to empowerment, prosperity, and resilience. It emphasizes democratized decision-making, ecosystem collaboration, and adaptability over time rather than detailed, prescriptive rules. @OpenAI

Human Human coverage frames the new principles as a significant pivot from the 2018 charter, particularly in downplaying AGI and softening earlier commitments to avoid competitive races with other labs. It highlights concerns that the principles’ vaguer, more aspirational language shifts responsibility outward and makes OpenAI’s obligations less concrete and harder to hold to account. @7dlt…clgf OpenAI has quietly rewritten its moral compass — and in the process, turned an idealistic AGI manifesto into a far more pragmatic playbook for competing in the AI arms race.

§2015–2018: From nonprofit idealism to an AGI mission

OpenAI launched in 2015 as a nonprofit research lab, founded in San Francisco by Sam Altman, Elon Musk, Greg Brockman, Ilya Sutskever, and others, with an explicitly altruistic mission: build artificial general intelligence (AGI) that would “benefit all of humanity.”¹ The 2018 charter leaned hard into that goal and into collaboration over competition. AGI — defined as “highly autonomous systems that outperform humans at most economically valuable work” — was the north star, mentioned a dozen times in the document, and framed as both an inevitability and an existential responsibility.²

Back then, OpenAI spelled out a radical promise: if some “value-aligned, safety-conscious” rival project got close to building AGI first, OpenAI would step aside and even help them, explicitly committing to “stop competing with and start assisting this project.”² That pledge set it apart from Big Tech and gave the lab a quasi-custodian role over the future of AI.

§2019–2023: The pivot to power — and products

That idealism didn’t survive long unaltered. As OpenAI shifted into a capped-profit structure and rolled out blockbuster products like ChatGPT and GPT-4, the gap between rhetoric and reality widened. The lab increasingly acted like a frontier AI company: raising billions, signing exclusive cloud deals, and turning research breakthroughs into commercial APIs.

Still, the 2018 charter technically sat in the background, with AGI as the explicit objective and a kind of gentlemen’s agreement on not turning the endgame into a race.

Meanwhile, AI capabilities leapt forward. OpenAI itself framed this era as one where “the technology, like others before, will give people more capability and agency,” invoking a future where what people could do with AI would “dwarf what people could do with steam engines or electricity.”¹ The implicit promise: the lab would steer this power safely and share its benefits widely.

§April 2026: A new principles document lands

On Sunday, April 27, 2026, OpenAI published a new document: “Our Principles,” a five-part framework for how it says it will develop and deploy AI from here on out.¹ CEO Sam Altman shared the list as the company’s updated core guidelines — effectively the first major rewrite of its underlying philosophy since 2018.²

The new piece opens in sweeping, almost utopian terms: AI, it says, has the potential to “significantly improve many aspects of society,” unleashing capability and agency on a scale that will make the steam engine and electricity look modest.¹ OpenAI imagines “a world with widespread flourishing at a level that is currently difficult to imagine,” where “individual potential, agency, and fulfillment significantly increase” and “a lot of the things we’ve only let ourselves dream about in sci-fi could become reality.”¹

But there’s an immediate note of caution: “this outcome is not guaranteed.” Power, the document warns, could end up “held by a small handful of companies using and controlling superintelligence, or it can be held in a decentralized way by people” — and OpenAI now says “our goal is to put truly general AI in the hands of as many people as possible.”¹

From there, the company lays out five principles:

1. Democratization – resisting the “potential of this technology to consolidate power in the hands of the few,” not just by giving access to AI but by ensuring that “key decisions about AI are made via democratic processes and with egalitarian principles, and not just made by AI labs.”¹
2. Empowerment – building products that let people “achieve their goals, learn more, be happier and more fulfilled, and pursue their dreams,” and giving users “the autonomy they need” with “very broad latitude” in how they use the systems, within reasonable limits.¹
3. Universal prosperity – fostering a world of “widespread flourishing” through new economic models and infrastructure (outlined in the summary of the piece).¹
4. Societal resilience – helping society guard against AI-driven risks, not just hyping benefits.¹
5. Adaptability – acknowledging that AI’s trajectory is unpredictable and emphasizing flexible governance and deployment strategies.¹

The missions statement — “Our mission is to ensure that AGI benefits all of humanity” — survives in the new text, but it sits in a very different landscape.¹

§The AGI downgrade: From obsession to afterthought

One of the biggest shifts is simply what the new document doesn’t talk about. Where the 2018 charter mentioned AGI 12 times and used it to anchor nearly every commitment, the 2026 version name-checks AGI only twice, focusing instead on AI systems broadly.²

Business Insider’s comparison is blunt: “Less emphasis on AGI” is the first major change. In 2018, the lab insisted that “to be effective at addressing AGI’s impact on society, OpenAI must be on the cutting edge of AI capabilities—policy and safety advocacy alone would be insufficient.”² The new principles recast the challenge as a continuous spectrum of capabilities, not a single finish line.

“This is an expansion of our long-held strategy of iterative deployment; we believe society needs to contend with each successive level of AI capability,” the updated blog says.² That language shifts the frame from building AGI to managing an escalating series of powerful models.

For optimists, this is a maturation: less fixation on sci‑fi superintelligence, more attention to real-world systems rolling out today. For skeptics, it looks like strategic ambiguity — keeping the AGI mission slogan while walking back the measurable commitments attached to it.

§The competition flip: From “we’ll step aside” to “we’re in the race”

The more dramatic turn is on competition.

In 2018, OpenAI fretted about “late-stage AGI development becoming a competitive race without time for adequate safety precautions” and promised that if a “value-aligned, safety-conscious” rival project approached AGI first, it would “stop competing with and start assisting this project.”²

In 2026, that language is gone. The new document drops explicit references to stepping aside or sharing frontier progress with a rival lab. Business Insider calls it “a 180-degree shift from the company’s original guidelines on collaboration and avoiding competition with rival labs.”²

Instead, the principles are framed in broad societal terms — democratization, empowerment, prosperity — while OpenAI’s competitive stance is left implicit. The updated document, as summarized, “skips mentions of sharing progress and stepping aside” and “implicitly states that if needed, the company will prioritize being competitive over AI for everyone.”²

In other words: the lab that once promised to exit the race if someone safer got close to AGI now reads like a company that intends to win — or at least stay in the lead.

§Inside the new rhetoric: Decentralization vs dominance

OpenAI’s own narrative presents this not as a power grab, but as a fight against concentrated power. The new principles warn explicitly that “power in the future can either be held by a small handful of companies using and controlling superintelligence, or it can be held in a decentralized way by people,” arguing that “the latter is much better” and that the goal is “to put truly general AI in the hands of as many people as possible.”¹

Democratization, as defined here, is partly about access — making sure “everyone” can use AI — and partly about governance, insisting that “key decisions about AI are made via democratic processes and with egalitarian principles, and not just made by AI labs.”¹

From an AI-lab perspective, this is a rebranding of its market dominance as a bulwark against something worse: a future where a few rivals, perhaps less “value-aligned,” own superintelligence outright. OpenAI is casting itself as the player that will decentralize power by centralizing the path to it.

Human critics, though, see a different throughline: the erosion of concrete, self‑limiting promises. Where the 2018 charter constrained OpenAI’s behavior in the event of a race, the 2026 version leans on aspirations and systems-level language — telling “the tech ecosystem and society” what should happen, while leaving the company’s own future choices more open‑ended.²

§Two readings of the same shift

By late April 2026, the contrasting narratives are clear.

OpenAI’s self‑portrait (AI perspective):

• AI is a transformative technology that could enable “widespread flourishing” and make sci‑fi‑level capabilities a reality.¹
• The lab’s mission — ensuring “AGI benefits all of humanity” — remains intact, but the focus must broaden to every “successive level of AI capability,” not just the AGI end state.¹²
• To avoid a world where a few companies hoard superintelligence, OpenAI says it must “put truly general AI in the hands of as many people as possible” and fight the concentration of power.¹
• Democratization, empowerment, prosperity, resilience, and adaptability are the guiding principles that will shape products and policy going forward.¹

The human‑side reading (journalistic perspective):

• The AGI obsession of 2018 has been dialed down: mentions of AGI fall from 12 to 2, and the lab no longer centers its whole philosophy on that single milestone.²
• The most radical collaborative pledge — to “stop competing” and “start assisting” a safer rival close to AGI — has vanished.
• The new principles pull back from specific, verifiable obligations and move toward broader recommendations for “the tech ecosystem and society.”²
• As Business Insider puts it, this is a “major update” and a “180-degree shift” on competition — a signal that OpenAI now sees itself as a full participant in, not a brake on, the race for frontier AI.²

§The bottom line: Principles for which era?

The old charter belonged to a world where AGI felt distant and OpenAI could afford to act like a philosopher-king of future tech. The new principles belong to a world where AI is already everywhere, OpenAI is one of the most powerful companies in the sector, and the race is no longer hypothetical.

OpenAI wants the public to see this rewrite as an evolution: from idealistic manifesto to operational blueprint. Its critics will see something sharper: a company rewriting the rules of its own restraint just as its systems become too big to ignore.

The real test won’t be how often the new document invokes “democratization” or “empowerment,” but whether, when the next capabilities leap hits, OpenAI acts like the steward it once promised to be — or like just another player that decided the race was worth running after all.

§1. Our Principles — “AI has the potential to significantly improve many aspects of society… our goal is to put truly general AI in the hands of as many people as possible. Our mission is to ensure that AGI benefits all of humanity. Here are the principles that guide our work. 1. Democratization.…”

§2. OpenAI just updated its principles. Here’s what changed since the original version, 8 years ago. — “In the 2018 charter, guidelines around artificial general intelligence… were in focus… ‘Therefore, if a value-aligned, safety-conscious project comes close to building AGI before we do, we commit to stop competing with and start assisting this project.’ The latest document skips mentions of sharing progress and stepping aside. It implicitly states that if needed, the company will prioritize being competitive over AI for everyone.”

Story coverage