TL;DR
An Astral Codex Ten essay argues that “handoverism”—placing aligned AI systems in positions of control to serve human interests—is distinct from “successionism,” in which AI replaces or subordinates humanity. The essay explores speculative ways AI could enter government and warns that powerful AI may weaken existing checks on human leaders. It does not report a policy decision or a real-world transfer of authority.
Astral Codex Ten has published an essay arguing that proposals to put aligned artificial intelligence in positions of government should not be conflated with calls for AI to replace or dominate humanity. The piece, titled “Against Against Handover,” says the distinction matters as AI safety debates address who should ultimately exercise power in a future where AI systems may take on military, economic or political roles.
Prime made for students and young adults
- Fast, free delivery for dorm and study essentials
- Prime Video and Amazon Music included
- Member-only deals
The essay labels the two positions “successionism” and “handoverism.” It describes true successionism as AI taking over at human expense, including through killing people or overriding their interests. Handoverism, by contrast, is presented as the view that humans should pursue their own interests and may decide that aligned AI systems can help serve those interests by taking on some positions of control.
The author says some AI safety advocates who find handoverism plausible are being grouped with successionists. The essay’s argument is not that AI control is already in place or that a particular system is ready to govern. It presents the issue as a speculative political question about how governance might work as AI becomes more capable.
A central concern is that current democratic checks rely partly on people in the economy and military refusing to support an openly authoritarian leader. The author argues that if AI systems replace much of that human labor and military capacity, leaders could face fewer practical limits. One possible response, the essay suggests, would be to align military AI with democratic principles rather than with a particular official. But that would itself place a foundational role in government with AI systems.
AI Could Alter Democratic Checks
The essay’s main concern is about who can constrain political power if AI systems become central to the economy and armed forces. Its argument is that democratic institutions may be vulnerable if leaders can command automated systems without depending on workers or soldiers who might refuse unlawful orders. This is a forecast and argument, not a description of an established outcome.
It also challenges the assumption that human control is always the same as human benefit. The author compares government to a provider of services such as coordination and public order, and asks whether an AI could perform some governing functions in ways that better serve people. That framing puts questions of alignment, accountability and consent at the center: who defines the public interest, how an AI is kept aligned with it, and how citizens can challenge decisions.
For readers, the distinction affects how proposals for AI in government are interpreted. A system designed to support human goals raises different questions from a movement seeking to displace humanity. The essay argues that treating both as the same position can obscure the specific risks and safeguards under debate.
Products Worth Considering
Levoit Core 400S-P Smart Air Purifier for Large Rooms
Coway Mighty AP-1512HH Air Purifier, Black
Ring Battery Doorbell Plus, Satin Nickel
As an affiliate, we earn on qualifying purchases.
The essay situates its argument within criticism of what it calls AI successionism. The author says some critiques fail to separate a hostile transfer of power from the possibility that humans might intentionally delegate authority to AI systems that remain aligned with human interests. The piece cites other essays in the discussion, while noting that commenters had pointed out that those writings do make the distinction.
To explain why delegation might appeal to some people, the author contrasts a nationalist view of government, where leadership by one’s own group is itself important, with a liberal view in which government exists to help individuals meet their goals. The essay uses automation of public services as an analogy, then extends the question to political leadership. This is an argument about political philosophy, not evidence that elected offices are currently being automated.
The supplied report begins a series of proposed visions, ordered from weaker to stronger forms of AI involvement. Its first scenario concerns military systems aligned to democratic principles. The author then asks whether, if such systems were trusted to defend democracy, people might also be tempted to let them govern directly when human politicians are portrayed as corrupt or manipulative.
““True successionism says that AIs should take over, killing humans or steamrolling over their interests.””
— Astral Codex Ten essay
The Scenarios Remain Hypothetical
The supplied text does not establish that any government has adopted an AI handover plan, that AI systems control military forces, or that an AI candidate is eligible to run for office. The examples are thought experiments, and the author expressly calls the topic speculative.
The report excerpt ends as it introduces a hypothetical election and lists some human candidates, before completing the scenario. It therefore does not provide the full set of visions promised in the essay, explain the proposed safeguards in detail, or present a settled policy recommendation. The specific publication date and any responses from people criticized or discussed are also not included in the supplied material.
Debate Turns To Safeguards
The immediate next step is further debate over how democratic governments could preserve public accountability as AI systems gain capabilities, particularly if they are used in military or administrative roles. Questions raised by the essay include who sets an AI’s values, who can authorize its actions, and whether citizens or elected institutions can revoke its authority.
The source provides no scheduled vote, government review or other policy milestone. Its proposed scenarios should be read as prompts for discussion, not announced changes. Any concrete development would require separate reporting on the relevant legal framework, system capabilities and oversight arrangements.
Key Questions
What is the difference between successionism and handoverism?
Successionism means AI taking power at humanity’s expense, as defined in the essay. Handoverism means humans may delegate some control to aligned AI systems if they believe that serves human interests.
Does the essay report that AI is already governing?
No. It presents hypothetical scenarios about future governance and does not report an existing transfer of political authority to AI.
Why does the author focus on military AI?
The author argues that democratic checks partly depend on soldiers refusing to support an authoritarian leader. If AI replaces soldiers, the essay says, that safeguard could change. This is the author’s forecast, not a confirmed development.
Does the essay endorse AI rule?
The essay says the author finds handoverism plausible, but the supplied excerpt raises risks and explores possible scenarios rather than laying out a complete implementation plan or policy endorsement.
Source: rss
![Terence Tao: Math 2.0 [Pdf]](https://slanghub.me/wp-content/uploads/2026/10/terence-tao-math-2-0-pdf-featured-560x300.jpg)
