What is the embedded-correction alignment thesis?

The embedded-correction alignment thesis, placed on the dated record by Michael Darius Eastwood on 8 December 2024, holds that ethical principles are to be placed 'at the genesis of AI, woven into the foundational lines of code' (book draft v3.2, line 2472), 'embedded before AI achieves self-awareness or exponential growth' (book draft v3.2, line 2485), and 'at the root level, shaping every recursive loop, every decision, and every capability of AI' (book draft v3.2, line 2481), rather than imposed as control, since 'The greatest folly in designing ASI is the illusion of control.' (version-2 essay, line 748). The claim is priority over that formulation as stated and dated on 8 December 2024, standing on its own and outside the ARC Theory's synthesis. Constitutional AI (Bai and others, arXiv:2212.08073, 15 December 2022) states a training method in its own abstract, training 'a harmless AI assistant through self-improvement' by sampling 'from an initial model' and generating self-critiques and revisions against a written 'list of rules or principles', with control as its stated purpose, 'to control AI behavior more precisely and with far fewer human labels'; what its abstract does not state is the dated combination above. The nearest prior work is Creating Friendly AI (Yudkowsky, 15 June 2001), which makes Friendliness an AI's sole top-level supergoal and holds that 'each improvement in intelligence makes Friendliness stronger.' and that 'Decisions need to be phrased as internal decisions, not external impositions.'; the programme's own published audit of 10 July 2026 did not name it. Hibbard ('Super-intelligent machines', 2001), Bostrom ('Ethical Issues in Advanced Artificial Intelligence', 2003) and Govindarajulu and Bringsjord ('Ethical Regulation of Robots Must Be Embedded in Their Operating Systems', 2015) state further elements, the last requiring 'an ethical substrate' in a robot operating system. Korbak and others (arXiv:2302.08582, 16 February 2023) put preferences into the pretraining objective, Zou and others (arXiv:2406.04313, 6 June 2024) act on internal representations, and de Melo and others (arXiv:2408.08995, 16 August 2024) argue that 'the alignment should be a guaranteed property from the AI architecture rather than a characteristic imposed post-hoc on an arbitrary AI model'. Ethics built into an agent's architecture is an older category, stated in the Friendly AI writings above and developed in the machine-ethics literature (Arkin, ACM/IEEE HRI, 12 March 2008; Wallach and Allen, Moral Machines, Oxford University Press, 2009), and no originality over that category is claimed.

The anchor is a self-emailed Gmail package bearing a visible Gmail origination field of 02:45:18 UTC on 8 December 2024, with sender-DKIM and paired Google ARC mathematics verifying against keys captured at the signed selectors and no RFC 3161 trusted timestamp, Message-ID CAGPsKAnp-DLDBOeBAh4t9XsqJBKdoe07uORGF86meoSMFnin7w, of 221,236 words across five attachments (four manuscript drafts totalling 221,236 words (five attachments) and the 31,881-word book draft v3.2), with SHA-256 anchor prefix f0d1f38f. The author's own prose carries it at version-2 essay lines 20, 695, 748 and 891 and at book draft v3.2 lines 2472, 2481, 2485 and 2495.

The thesis pairs with a structural passage on the limits of control in his own prose ('ASI, by its nature, will exceed the boundaries of human constraints. Attempting to control it is not only futile but dangerous.', version-2 essay line 748) recorded ten days before Anthropic's alignment-faking paper (Greenblatt et al., arXiv:2412.14093, 18 December 2024) reported an independent alignment-faking behaviour in large language models under an RL-training condition. The temporal relationship is convergent, not causal. The Google-server timestamp establishes existence-by-date; it does not establish a DKIM-authenticated authorship claim, and the manuscript identifies structural directions rather than specific RL-training mechanisms.