OpenAI compares AI to electricity and announces its third phase

What happened?

OpenAI has published a manifesto signed by its CEO, Sam Altman, drawing a historical parallel between artificial intelligence and electricity. Altman argues that both technologies transform society gradually but profoundly, and announces the beginning of a 'third phase' in the company's evolution. This new stage follows the phases of foundational research and product deployment like ChatGPT, and will focus on 'building technology to benefit everyone,' with an emphasis on collaboration and shared governance. The manifesto was published on February 12, 2025, according to TechRadar, and aims to establish a long-term vision for the company.

Why is it important?

The comparison with electricity is not accidental. Electricity took decades to transform the economy: in 1920, only 35% of U.S. households had electricity, and by 1950 that figure exceeded 90%. During that period, new industries such as home appliances, electronics, and computing emerged, and millions of jobs that did not exist before were created. By using this analogy, OpenAI seeks to calm fears about mass job displacement and position AI as basic infrastructure that will enhance human capabilities. Moreover, the 'third phase' implies a shift in focus: from selling APIs and chatbots to building ecosystems where AI is a regulated public good. This could influence how other tech companies, like Google (with Gemini) or Meta (with Llama), chart their roadmaps. In fact, Meta has already bet on a more open model with its open-source Llama models, while Google maintains a more controlled approach. OpenAI's manifesto could be an attempt to balance its public image at a time of increasing global regulation.

The regulatory context is key. The European Union approved the AI Act in 2024, which classifies AI systems by risk level and requires transparency for general-purpose models. In the United States, the 2023 executive order on safe and trustworthy AI sets testing and reporting standards. Altman's manifesto arrives just as regulators seek to define how to govern AI. By stating that 'automating everything is not the future we want,' Altman distances himself from the more radical 'automation everything' narrative that has dominated the discourse of some investors, like Marc Andreessen, who has advocated for accelerated automation. This stance could help OpenAI gain legitimacy with regulators and the public, but it also reflects an internal tension in the company: while Altman promotes a collaborative vision, OpenAI has been criticized for its secrecy and for moving away from its nonprofit origins.

Consequences and analysis

The manifesto comes at a time of increasing global regulation and public skepticism. According to a 2024 Pew Research survey, 52% of Americans feel more concerned than excited about the impact of AI. However, the document offers no concrete details on how this vision will be implemented, leaving room for speculation. Analysts point out that it could be a public relations move to gain legitimacy with regulators, or a real indication that OpenAI is willing to cede some of its centralized control in favor of more open governance. The company has already taken steps in that direction: in 2024, OpenAI launched a $5 million fund for AI governance research and has participated in forums like the UK's AI Safety Summit. Nevertheless, the manifesto omits mentioning risks of bias, job displacement, and power concentration, which has been criticized by tech ethics experts like Timnit Gebru, who has noted that promises of 'benefit for all' often hide corporate interests.

The market impact could be significant. If OpenAI truly moves toward a more open and collaborative model, it could pressure competitors to follow suit. For example, Anthropic, the rival startup founded by former OpenAI employees, has already adopted a 'constitutional AI' approach and published documents on its safety principles. However, the lack of concrete details in Altman's manifesto makes it difficult to assess whether this is a real change or a communication strategy. Investors will also be watching: OpenAI is valued at $150 billion after its latest funding round in October 2024, and any sign that the company prioritizes public good over profitability could affect its valuation. On the other hand, if the manifesto is merely a symbolic gesture, it could increase skepticism toward the company.

From a historical perspective, the analogy with electricity is powerful but imperfect. Electricity did not pose existential risks or control problems of a technology that could surpass human intelligence. Moreover, the adoption of electricity was driven by private companies and governments, but there was no global debate about its governance like the one we see today with AI. OpenAI's manifesto seems to ignore these nuances, which could be seen as an oversimplification by academics and regulators.

What readers should know

For professionals and businesses, the key message is that OpenAI is betting on an AI-as-public-service model, not as a job replacement. This suggests that investments in training and adaptation will be more important than ever. AI could automate up to 30% of work tasks in advanced economies by 2030, according to McKinsey, but it will also create new roles in areas like ethics, governance, and system maintenance. Developers should expect more open APIs and possibly open-source tools, although OpenAI has not yet confirmed concrete plans. Regulators will find in the manifesto an argument to push for collaborative governance frameworks, like those already discussed at the 2025 Paris AI Summit. However, skepticism is healthy: it remains to be seen whether OpenAI's actions align with its words. Meanwhile, companies and workers should prepare for an environment where AI will be ubiquitous, but its governance is still under construction.