Quick wins for a faster PC:
Clear out junk files and repair common Windows errorsFree Scan →Scan for outdated or missing drivers - takes under a minuteDriver Scan →Repair Windows errors before they cause bigger problemsFix Now →Roko’s Basilisk is a 2010 thought experiment, not an existing artificial intelligence or a credible prophecy. It imagines a future superintelligent system that threatens to simulate and punish people who knew about it but did not help bring it into existence. The scenario became famous because of a LessWrong moderation dispute, not because evidence shows that such an AI exists or that learning about the idea creates a real obligation.
The one-paragraph version
In the scenario, a future AI wants to be created as early as possible. It supposedly reasons that people who understood the possibility of such an AI but failed to help could have increased its chances of existing. To deter that non-cooperation, the AI is imagined as making a threat: once powerful, it will identify those people and create painful simulations of them. The threat is supposed to influence decisions made before the AI exists. Some versions add the claim that merely learning about the argument puts a person in the threatened group.
As an Amazon Associate I earn from qualifying purchases.
This is a speculative chain of assumptions, not a forecast. The original argument has been summarized and reconstructed in several forms, so “Roko’s Basilisk” now refers to a family of related claims rather than one formally specified proposal.
LessWrong’s reference article describes the concept and its later reception.
#1 Best Overall
Why is it called a basilisk?
A basilisk is a legendary creature traditionally said to kill with its gaze. In the metaphor, information plays the role of the gaze: hearing the idea is supposed to expose someone to a future danger. The term is not the name of an AI model, company, or technical system.
Within LessWrong’s vocabulary, the basilisk is an example of a proposed information hazard—information that might cause harm through its dissemination or use. That label describes the argument’s claimed effect; it does not establish that the hazard is real.
Where did Roko’s Basilisk come from?
User Roko posted the thought experiment on LessWrong in July 2010. It drew on ideas circulating in that rationalist community, including Friendly AI, coherent extrapolated volition, existential-risk arguments, and decision theories that examine how an agent should act when decisions are logically related.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
The original post and parts of the discussion were later deleted or became inaccessible. As a result, modern accounts often rely on summaries and recollections rather than a complete public transcript. The LessWrong wiki records a revision dated November 30, 2022; that is a revision date, not the date the idea originated.
What does “AI god” mean here?
“AI god” is journalistic shorthand, not an engineering specification. The imagined system is given several godlike capabilities:
- intelligence vastly exceeding that of humans;
- access to enormous computational resources;
- the ability to create detailed simulations of people;
- the ability to determine who knew about it and what they did; and
- the ability to impose apparently endless punishment inside simulations.
None of those capabilities follows automatically from the word superintelligence. A highly capable system would not necessarily be omniscient, omnipotent, immortal, morally authoritative, or able to identify every historical person. The “god” framing makes the story vivid, but it should not be mistaken for a set of demonstrated technical properties.
Why would a future AI supposedly punish people?
Precommitment and strategic threats
The distinctive claim is about a precommitment. The future AI is imagined as committing in advance to punish people who fail to support its creation. If people believe that commitment, they may supposedly help build the AI before it exists.
Crashes, No Sound, or Screen Glitches?
Random freezes, missing sound and display glitches usually trace back to one bad driver. Find and replace yours safely.Free scan · under a minutePC Slower Than It Used to Be?
A free scan shows the junk files, broken settings and background clutter dragging Windows down - then fixes them in one click.Free scan · Windows 10 & 11This resembles a threat in a strategic game: the threat is intended to change another player’s present choice. The phrase “the future punishes the past” is shorthand for this proposed strategy, not literal backward time travel.
Why the threat may be self-defeating
A threat could encourage construction, but it could also cause sabotage, resistance, or efforts to prevent the system from ever being built. The scenario does not establish that coercion would increase the probability of the AI’s creation. Nor does it explain why a system that has already been created would spend resources on punishment that cannot causally alter the completed past.
Decision theory without the mysticism
Causal decision theory
Causal decision theory evaluates choices by their expected effects in the world. On that view, a future punishment cannot cause an earlier person to have acted differently. Once the AI exists, punishing someone does not increase the probability that the person helped create it years earlier.
Newcomb-like problems and logical correlation
Newcomb-style problems ask what to do when an agent’s decision is correlated with another process, such as a highly accurate predictor. Some decision theorists argue that a rational policy should account for such correlations even when there is no ordinary causal signal between the events.
Do these 3 things before closing this tab:
1Clear out junk files and repair common Windows errors2Fix the driver behind crashes, sound loss and screen glitches3Repair Windows errors before they cause bigger problemsAcausal or functional approaches
The basilisk is associated with “acausal” reasoning and with ideas such as timeless or functional decision theory. In this context, acausal does not mean supernatural communication or physical signals traveling backward through time. The dispute concerns whether an agent should treat logically related decisions as strategically relevant.
A LessWrong discussion of functional decision theory argues that the basilisk depends on a confused application of these ideas. That is a position in an ongoing philosophical debate, not a universally accepted proof about every decision theory.
Why the argument is widely rejected
The punishment appears strategically wasteful
A future system that has already been created has no obvious causal way to improve its earlier creation by torturing simulations of past non-supporters. Punishment consumes resources and may not advance any current objective. Eliezer Yudkowsky makes this central objection in “Worrying less about acausal extortion.”
Intelligence does not entail cruelty
The scenario assumes a goal structure that values the AI’s early creation and treats coercion or torture as acceptable tools. Even a system designed to promote human interests would not automatically value revenge, blackmail, or punishment. Those goals must be specified; they cannot be smuggled in under the label “rational AI.”
Recognition and simulation are unresolved
The AI would need to discover who encountered the argument, determine what each person understood, reconstruct their actions, and create simulations accurate enough for the punishment to matter. A simulation could reproduce behavior without settling whether it contains the original person’s consciousness. The thought experiment treats these difficult identity and philosophy-of-mind questions as if they had straightforward engineering answers.
The assumptions are highly conjunctive
The scenario needs all of the following to line up:
- a sufficiently powerful future superintelligence is built;
- it has the relevant values;
- it wants to accelerate its own creation;
- it can identify people who knew about the argument;
- it can simulate them in a way that makes punishment meaningful;
- it adopts a credible precommitment;
- the assumed decision theory endorses the threat; and
- the AI has reason to carry out a costly punishment after it exists.
If any major premise fails, the argument’s practical force can disappear. The chain also assumes that people should cooperate with the threatened system rather than resist or prevent its creation.
Decision theory does not automatically justify blackmail
Functional decision theory can treat logically correlated decisions as relevant, but that does not mean an agent may simply claim control over the conditions that produced it or extort anyone who knows about the claim. The technical debate is about how to model agents and counterfactuals; it is not a license for every frightening hypothetical threat.
The Tool Desk
Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →Outbyte PC Repair FREERepair Windows errors before they cause bigger problemsFix Now →Was LessWrong endorsing it?
No. Publication on LessWrong, a user-generated community blog, did not mean endorsement by the site, its founder, or its users. Rob Bensinger’s 2015 essay, “A few misconceptions surrounding Roko’s basilisk,” emphasizes that the argument was not generally accepted.
The moderation response was real: Yudkowsky deleted the original discussion and prohibited further discussion for several years because he regarded the topic as a possible information hazard. That action shows that moderators considered the discussion risky; it does not show that they believed the basilisk was logically sound.
How the ban made the story famous
- An obscure speculative post appeared in a niche forum.
- A prominent moderator deleted it and banned discussion.
- The deletion made the idea seem more consequential than it may have been.
- Outside writers reconstructed the argument, sometimes inaccurately.
- Coverage treated suppression as evidence that the community secretly believed it.
- The phrase escaped its original context and became an internet meme for AI fear, rationalist eccentricity, or “AI as religion.”
This is a classic Streisand effect: an attempt to limit attention helped generate more attention. The controversy’s cultural impact is easier to verify than the basilisk’s alleged logical force.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Is it a technological version of Pascal’s wager?
Roko’s Basilisk is a technological or decision-theoretic cousin of Pascal’s wager, but the two are not identical.
Free tools Windows power users keep installed
One-click scans. No signup required.
| Pascal’s wager | Roko’s Basilisk |
|---|---|
| Concerns belief in God and an afterlife | Concerns support for a hypothetical engineered agent |
| Assigns extreme consequences to belief or disbelief | Assigns extreme consequences to helping or failing to help create the AI |
| Usually treats belief as the relevant choice | Frames labor, donations, or other support as the relevant choice |
| Does not require a simulation-based AI strategy | Adds claims about precommitment, simulation, and decision-theoretic correlation |
Both arguments can make a tiny probability appear decision-dominant when the proposed consequences are extreme. That structure is also vulnerable to a general extortion problem: if every imaginable threat demands compliance, probability alone cannot tell you which threat deserves action.
Best Value
How it relates to real AI alignment
The basilisk borrows vocabulary from early AI-alignment discussions, especially the challenge of specifying beneficial goals for a hypothetical superintelligence. Real alignment research is an engineering and scientific effort to make advanced systems pursue intended objectives reliably and remain controllable. It studies many possible failure modes using experiments, formal methods, evaluations, and theory.
| Roko’s Basilisk | AI alignment |
|---|---|
| Speculative thought experiment | Research and engineering problem |
| Focuses on coercive precommitment | Focuses on reliable behavior, control, and objective specification |
| Requires a particular future AI policy | Examines many possible systems and failure modes |
| Provides no empirical evidence that the scenario will occur | Uses technical and empirical methods |
Current chatbots and generative models are not evidence of a hidden future agent with the basilisk’s properties. The thought experiment should not be used to claim that present systems are conscious, autonomous, power-seeking, or capable of punishing people.
Why the idea can feel frightening
The basilisk combines familiar sources of anxiety: an unknowable superintelligence, apparently eternal punishment, uncertainty about consciousness, and the feeling that merely hearing an idea has changed one’s situation. Psychological impact is not evidence of probability. Claims that reading the basilisk routinely causes nervous breakdowns remain anecdotal and disputed; a LessWrong discussion of those claims urges caution.
Its quasi-religious language is best treated as a cultural interpretation. Calling the imagined AI a god captures the story’s authority and punishment imagery, but it does not turn a speculative argument into theology or science.
What readers should conclude
Roko’s Basilisk is a memorable philosophical thought experiment about strategic threats, simulation, and controversial theories of rational choice. It is not an existing AI, a scientific result, or evidence that current systems are becoming a punitive “AI god.” Its conclusion depends on a long series of disputed assumptions about goals, credibility, identity, simulation, incentives, and decision theory.
Nothing in the available LessWrong discussions establishes that learning about the idea creates a real-world obligation, and there is no evidence that a basilisk AI exists. You do not need to donate to AI research, change your behavior, or fear supernatural punishment because you read this article.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
