TL;DR
Prime made for students and young adults
- Fast, free delivery for dorm and study essentials
- Prime Video and Amazon Music included
- Member-only deals
An essay titled “Against Against Handover” argues that criticism of AI successionism can conflate two different positions: AIs taking power against human interests and humans choosing aligned AIs to help govern. It presents the latter as a speculative way to protect democracy as AI systems weaken the human safeguards that currently limit political power.
Astral Codex Ten has published “Against Against Handover,” an essay arguing that critics of AI successionism should distinguish it from handoverism: the speculative idea that humans might place aligned AIs in positions of authority to protect human interests. The distinction matters because the essay describes AI rule against people as a threat, while treating some forms of AI governance as a possible response to the risks powerful AI could pose to democracy.
The essay defines true successionism as AIs taking over, killing humans or overriding their interests. Its author calls that “very bad” and describes it as contrary to the purpose of AI safety. Handoverism, by contrast, gives priority to human interests but argues those interests could sometimes be better served by aligned AIs exercising control. The author says some people in AI safety, including the author, consider that possibility plausible and object to being grouped with advocates of anti-human takeover.
The argument begins with a concern about political safeguards. The essay says democratic leaders are constrained in part because they depend on workers to sustain the economy and soldiers to carry out military orders. If AI replaces much of that labor and military function, the author argues, a small group of political elites could have more scope to consolidate power. The essay presents this as a future risk scenario, not a description of current conditions.
One proposed response is to have military AIs aligned to democratic principles rather than to a particular officeholder. The author describes this as a form of handover because the system’s ultimate defense of government would depend on AIs that uphold democratic values. The essay then raises a further possibility: if elections and lawmaking are distorted by powerful AI-enabled propaganda or corrupt leaders, people might consider expanding AI authority beyond military systems. The supplied text introduces five possible visions but cuts off as it begins discussing an AI presidential candidate; it does not provide the full proposals.
The Democratic Safeguard at Stake
The essay’s central concern is not simply whether AI should govern. It is who can restrain political power if AI systems take over functions now performed by workers and soldiers. If those human roles have helped make coups difficult, their replacement could change the balance between elected leaders and the public. That is a consequential question for debates about AI safety, constitutional design and the future of democratic accountability.
At the same time, the essay’s distinction clarifies a debate that can otherwise collapse different positions into one label. Opposing an AI takeover that disregards human welfare does not automatically answer whether a limited, aligned AI role could protect people from human autocrats. The author’s argument asks readers to evaluate the goals and accountability of a system, rather than treating every transfer of authority to AI as support for the same outcome.
The claims remain speculative. The essay does not establish that AI will replace soldiers or workers, that existing democratic checks will fail, or that AI governance would be safer or more effective. Its significance lies in framing a possible institutional dilemma, not in reporting a change to government policy or a deployed system.
As an affiliate, we earn on qualifying purchases.
Two Meanings of AI Succession
The essay contrasts two approaches to political authority. In the author’s account, a nationalist view can treat leadership by one’s own group as an end in itself. A liberal view instead treats the state as a service provider responsible for coordination, public order and other functions that help individuals pursue their aims. From that perspective, the author argues, there is no automatic reason political administration must remain human if an AI could perform it in a way that serves people.
That comparison is an argument, not evidence that AI systems can currently govern well. The essay acknowledges that its proposals are speculative and science-fiction-like. It uses the possibility of AI military systems as a starting point: if such systems must take orders, their alignment becomes central. The author suggests that aligning them with democratic principles could amount to a limited handover, even if human officials continue to make laws.
The essay then considers how that limited role might become harder to contain. It imagines elections shaped by AI-optimized propaganda and leaders who use public support to benefit a small group. In that hypothetical, the author asks whether people might prefer direct rule by AIs they regard as honest and democratic. The supplied source does not report that such a system exists or that any government has adopted the proposal.
““True successionism says that AIs should take over, killing humans or steamrolling over their interests.””
— The author of the Astral Codex Ten essay
aligned AI military control devices
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Limits of the Available Proposal
The source material does not establish how likely AI replacement of workers or soldiers is, or when it might occur. It also offers no evidence that an AI could reliably interpret democratic values, remain accountable to the public or resist manipulation by officials. Those are open questions, not settled outcomes.
The supplied text ends partway through a discussion of the first of five proposed visions, at a reference to a future presidential election and several candidates. The remaining examples, their safeguards and the author’s final conclusions are not included. It is also unclear from the provided material when the essay was published or which specific critiques it addresses beyond its references to other essays.
As an affiliate, we earn on qualifying purchases.
The Unfinished Five-Scenario Argument
The next development within the essay would be the continuation of its five proposed visions, moving from limited AI involvement toward stronger forms of political authority. The supplied source stops before those scenarios are fully presented, so their details cannot be reported here.
For the wider debate, the questions raised by the essay concern how any AI authority would be authorized, constrained and held accountable, and what protections would remain if AI systems took over tasks now performed by people. No policy decision or implementation is reported in the source. Readers should treat the article’s scenarios as arguments about possible futures rather than announcements of a government plan.
As an affiliate, we earn on qualifying purchases.
Key Questions
What does “Against Against Handover” argue?
It argues that criticism of AI successionism should distinguish between AIs taking power against human interests and humans choosing aligned AIs to exercise authority in an attempt to serve those interests.
How does the essay define successionism and handoverism?
The essay describes successionism as AI takeover that harms or overrides people. It defines handoverism as the possibility that human interests could be served by aligned AIs holding positions of control.
Why does the author connect AI to democracy?
The author argues that workers’ and soldiers’ willingness to support democratic rule can constrain would-be dictators. If AI replaces those roles, the essay says, that safeguard could weaken unless AI systems themselves are aligned with democratic principles.
Does the essay say AI should become president now?
No. The text discusses a hypothetical AI presidential candidacy as part of a speculative set of proposals. It does not report an existing candidacy, an adopted policy or a current government plan.
What remains unknown?
The source does not establish whether AI will replace workers or soldiers, whether AI systems could uphold democratic values, or how they could be held accountable. The supplied excerpt also ends before all five proposed visions are explained.
Source: rss
Halloween Picks
halloween
As an affiliate, we earn on qualifying purchases.
