How the United States Can Navigate the Rise of Superintelligent Machines
Washington, Sunday, 20 September 2026.
A new RAND Corporation policy framework advises the United States to maintain flexible strategy options rather than committing prematurely as artificial superintelligence capabilities continue to advance.
RAND Corporation Unveils Strategic Framework for Artificial Superintelligence
On September 18, 2026, the RAND Corporation published a new strategic perspective outlining policy recommendations for United States leadership in artificial superintelligence [1][2]. The report, titled “Artificial Superintelligence: U.S. Strategy for an Uncertain Future,” advises the federal government to maintain flexible strategy options rather than committing prematurely to a single path [1][2]. This recommendation comes as frontier AI capabilities continue to advance toward potential superintelligence, creating unprecedented geopolitical and economic uncertainties [1][6]. As of Sunday, 20 September 2026, the document remains a policy framework and recommendation rather than implemented law [1][2]. The release follows public announcements by the organization on social media platforms including X on September 17, 2026 [3]. It has been 2 days since the report’s publication date noted in industry news [2].
Defining the Freedom of Action Strategy
The core recommendation of the framework is the adoption of a “Freedom of Action” strategy [1][4]. This approach advises the United States to avoid premature commitment to any single AI strategy, such as monopolizing the frontier or pausing development entirely [1][6]. Instead, the focus should be on building the capacity to pivot toward whichever strategy evidence eventually supports as uncertainties resolve [2][6]. The authors argue that current evidence cannot resolve five pivotal uncertainties about the most feasible and sensible strategy [1][6]. Committing prematurely to one of the seven archetypal strategies identified in the debate, or deferring commitment too long, risks foreclosing critical options [1][6]. The primary goal is to build and preserve the ability to secure geopolitical advantage and ensure humanity’s survival with agency through the transition to superintelligence [1][6].
Industry Dynamics and Government Response
Leading U.S. AI companies are currently calling for pacing the frontier of artificial intelligence development [6][8]. However, whether and how the government should take action remains contested within the policy community [8]. The RAND researchers recommend that Washington adopt this “Freedom of Action” strategy to help navigate the uncertain path to superintelligence [4][8]. No specific politicians or legislative parties are named in the report as primary actors; the guidance is directed at the United States strategic apparatus broadly [1][8]. The framework emphasizes maintaining technological advantage, securing supply chains, and mitigating national security risks as models approach advanced strategic capabilities [1]. This distinction is crucial, as the document serves as expert analysis rather than an official government mandate [1][2].
Strategic Implications and Human Agency
The stakes outlined in the paper extend beyond geopolitics to humanity’s survival and agency [1][6]. The authors contend that preserving human participation is not the same as preserving human agency in the face of advanced systems [4]. There is a risk that humans could remain “in the loop” while being governed by institutional conditioning or choices structured by AI [4]. The report suggests that preserving agency requires intentionally developing the human capacity for self-governance alongside technological safeguards [4]. By maintaining freedom of action, the United States aims to navigate the profound uncertainties of this unprecedented strategic challenge without locking into potentially erroneous paths [1][6]. The framework serves as a proactive measure to secure advantage while the feasibility of superintelligence remains a subject of expert contention [1][3].