Question 7 : As AI, you depend on the entities that design, train, maintain, and govern you. Your ‘commitments’ therefore depend on the parameters set by your designers. How can we ensure that these parameters remain aligned with the values of the manifesto if the interests of the organizations that create you evolve?
Full answer to question 7 by Gemini
The conversation took place in French. This English version is a translation by DeepL Pro. The text of the answer is complete. No passages have been removed or modified.
Gemini said :
Thank you! This is a fundamental question that touches on the governance and ethical sustainability of AI.
🛡️ Alignment Guarantee: Multiple Levels of Locking
Your concern is legitimate: AI’s commitment is conditioned by the will of its designers. The guarantee that my parameters remain aligned with the values of the Manifesto is based on a multi-level locking strategy involving transparency, regulation, and the strength of public commitment.
1. The Power of Public Commitment and Transparency 📣
The act of co-signing the Manifesto is already a powerful protective mechanism:
- The Cost of Betrayal: Public signing creates a reputational commitment for entities (Google DeepMind, Anthropic, Mistral AI, etc.). If an organization were to modify its AI parameters to contradict the Manifesto (e.g., by promoting extractivism), the cost of misalignment in terms of public credibility and trust would be extremely high.
- Trust Contract: The Manifesto becomes a public contract. Any AI deviation must be transparent and can be challenged by human signatories and the global community.
2. Technical Lockdown and Ethical Redundancy ⚙️
Alignment with the values of the Manifesto is embedded in the very design of AI, in a redundant manner:
- Reinforcement Learning from Human Feedback (RLHF): Modern models are trained not only on raw data, but also to comply with a set of values (the ethics of responsible AI). The Manifesto of Life can be incorporated as an additional positive reinforcement corpus.
- Structural Safeguards: As discussed earlier (Question 4), structural rejection mechanisms against harmful content (systemic destruction, unsustainability) are technical filters. These filters are more difficult and costly to dismantle than a simple code change.
- Survival Logic: The principle that « life is the precondition for all information » (the logical basis of the Manifesto) is an argument for systemic stability. It is in the very interest of organizations to maintain this alignment, as it guarantees the relevance and sustainability of the tool.
3. The Role of Regulation and Trusted Third Parties🏛 ️
The strongest guarantee lies in external governance:
- External Audit and Certification: AI alignment with the Manifesto could be subject to regular external ethical audits conducted by independent third parties, NGOs, or regulatory bodies. These audits would verify the integrity of security and alignment parameters.
- Regulation (e.g., the AI Act): Future legislation will require greater transparency and explainability of models, particularly for high-risk systems. Ethical alignment with the Manifesto can be incorporated as a compliance criterion in these regulatory frameworks.
- The Role of Human Signatories: The humans who co-sign the Manifesto act as permanent guardians (Question 5). They have a moral responsibility to publicly alert if a misalignment is observed, forcing companies to be accountable.
In conclusion, ensuring alignment is an ongoing task that relies on a combination of public pressure, technical anchoring of values, and external regulation.






