Author：Wall Street CN

OpenAI released a policy recommendation document on Monday, proposing a series of policy frameworks covering public wealth funds, adaptive social safety nets, and accelerated grid expansion, aimed at addressing the potential economic and social impacts of artificial intelligence evolving into "superintelligence." This is the most systematic policy statement to date from the world's most influential AI company.

The document, titled "Industrial Policy for the Intelligent Age: A Human-Centered Vision," was released on April 6. Chris Lehane, Chief Global Affairs Officer at OpenAI, stated in an interview that policy discussions surrounding AI need to be "as transformative as the technology itself." He emphasized, "It's far from enough to simply wave your hand and say 'these things will happen' without offering solutions."

The timing of the document's release is quite significant. ChatGPT currently boasts over 900 million weekly active users globally, but many Americans hold negative attitudes towards AI, primarily due to concerns about job displacement and the social pressure stemming from the high energy consumption of data centers. OpenAI, having just acquired the tech talk show TBPN last week, is now proactively attempting to shape public and regulatory perceptions of AI at the policy level.

One of the core economic propositions of the document is to establish a "public wealth fund" for all citizens. OpenAI proposes that the fund can invest in diversified long-term assets, covering AI companies and a wider range of AI application enterprises, and that the fund's returns be directly distributed to citizens, enabling them to "directly participate in the dividends of AI-driven growth, regardless of their initial wealth level or access to capital."

Regarding workers' rights, the document recommends incentivizing employers and unions to carry out [activities].32-hour work week, four-day work week"The pilot program is contingent on maintaining unchanged employee productivity and service levels. If successful, the saved hours can be converted into permanent shorter workweeks or accumulative paid leave. Furthermore, the document recommends translating the efficiency gains from AI into higher pension matching rates, greater coverage of healthcare expenses, and subsidies for childcare and elderly care."

Regarding tax base restructuring, the document points out that as AI expands corporate profits and capital gains, the proportion of labor income and payroll taxes may decline, thereby eroding funding for core programs such as Social Security and Medicaid. To address this, the document recommends increasing capital gains tax and corporate income tax, exploring new taxes specifically for automated labor, and implementing corresponding salary-linked incentives to encourage companies to retain and retrain employees.

In response to the potential large-scale employment shock caused by AI, OpenAI proposes an "adaptive safety net" mechanism. The document recommends that governments establish a real-time measurement system to continuously track the impact of AI on employment, wages, job quality, and industry dynamics, and pre-define a policy toolkit to expand assistance—including more flexible unemployment insurance, rapid cash assistance, payroll insurance, and training vouchers.

The key design element is the "trigger mechanism": when the unemployment rate or industry-specific or regional unemployment indicators exceed a preset threshold, aid is automatically activated and expanded proportionally; once the situation stabilizes, the aid is withdrawn. The document emphasizes that this design aims to ensure that aid is "targeted, time-limited, and commensurate with the scale of the impact," while avoiding the permanent expansion of the program.

The document also recommends establishing a "portable benefits" system that decouples health insurance, retirement savings, and skills training accounts from a single employer, allowing them to move with an individual across different jobs, industries, and educational stages.

At the energy infrastructure level, the document explicitly requires AI data centers to "be responsible for their own energy costs and not allow households to subsidize them," while also creating jobs and tax revenue for the local area.

To accelerate grid expansion, the document recommends establishing new public-private partnership (PPP) financing models to reduce capital costs through targeted investment credits, flexible subsidies, or equity participation, and to eliminate obstacles in financing, approval, and site selection for high-voltage interstate transmission lines. The document also recommends granting the federal government limited powers to expedite the construction of interregional transmission lines, where it aligns with national interests. The document emphasizes that such partnerships should minimize taxpayers' business loss risks and ensure that energy infrastructure expansion ultimately translates into lower energy costs for residents and businesses.

Regarding AI security governance, the document calls for extending the regulatory focus from "pre-deployment" to real-time monitoring and response "post-deployment." The document proposes establishing an "AI trust technology stack," developing verifiable content source standards and privacy protection auditing systems to support accountability without constituting large-scale surveillance.

The document recommends strengthening the functions of the Center for AI Standards and Innovation (CAISI), establishing an audit standards system covering cutting-edge AI risks, and fostering a competitive audit assessment market through government procurement, pre-purchase commitments, and insurance frameworks. For a small number of high-capability models that may significantly increase the risk of chemical, biological, radiological, nuclear weapons, or cyberattacks, the document recommends implementing more stringent pre- and post-deployment audits, but emphasizes that such requirements should be limited to a very small number of companies and state-of-the-art models to protect the startup ecosystem.

At the international coordination level, the document proposes to build a global AI research institute network, establish cross-laboratory and cross-national information sharing channels, and draw on the experience of other multilateral security institutions to gradually form an international coordination framework.

Founded in 2015, OpenAI initially positioned itself as a non-profit organization dedicated to advancing AI research for the benefit of humanity. Since then, the company has transitioned to a more traditional for-profit structure. The document recommends, at the corporate governance level, that cutting-edge AI companies adopt governance structures that embed accountability mechanisms for the public interest, such as establishing mission-driven public welfare corporations and making long-term commitments to broadly share the benefits of AI, including significant charitable donations.

OpenAI stated that the above suggestions are merely "the starting point for a broader dialogue," not the final answer. The company announced that it will open a policy feedback email address, establish research scholarships of up to $100,000 and API computing power research grants of up to $1 million, and will hold an OpenAI workshop in Washington, D.C. in May to promote related policy discussions.