Seven principles · one standard
Conviction carries responsibility.
These principles guide how humAIne researches, builds and publishes now—and set the standard for the operating and ownership work in its long-term thesis.
Read the principlesScope
humAIne is currently a one-human, pre-incorporation research venture. References to capital, boards and ownership express a future standard—not claims of a current portfolio, mandate or operating practice.
The principles
Clear enough to guide a decision.
Specific enough to test the work against. Strong enough to matter when the convenient choice points elsewhere.
Evidence before conviction.
Distinguish verified facts from estimates and judgement. Cite the sources behind material claims, test the case against contrary evidence and state where uncertainty remains. AI can widen the research surface; it cannot own the conclusion.
Ownership means accountability.
Capital carries duties as well as rights. The long-term standard is clear responsibility, fair treatment, relevant board expertise and incentives connected to durable performance. Influence should be specific, constructive and accountable for its consequences.
Human agency comes first.
Technology should increase people’s ability to decide, create, work and live well. Systems that can affect a person’s life need proportionate safeguards, meaningful human oversight and a practical way to challenge harmful outcomes.
Trust is part of the product.
Safety, privacy, security and honest communication belong in the product and business model from the start. Material uses of AI should be understandable, governed and tested for the ways they can fail or treat people unfairly.
People create the value.
People create and carry an organisation’s knowledge. Durable capability depends on fair treatment, learning and serious attention to how technology changes work—not only to the costs it removes.
Build value that lasts.
Climate, nature and resource constraints shape business economics. Where they are material, strategy should account for dependencies, real-world costs, resilience and the investment required to improve performance.
Integrity must be visible.
Reject bribery, corruption, concealed conflicts, manipulation and profit detached from real value creation. Protect confidential information, disclose relevant conflicts, correct material errors and walk away when clean work is impossible.
Principle into practice
The current work is where the standard starts.
A future promise is only credible if the research, tools and publishing choices already reflect it.
- 01
Source discipline
Material claims should be traceable to credible evidence, with estimates and inference labelled.
- 02
Visible authorship
Agent assistance does not obscure who formed the judgement and accepts responsibility for it.
- 03
Correction over pride
New evidence can change a thesis. Material errors should be corrected plainly rather than defended reflexively.
- 04
Proportionate care
Privacy, confidentiality and human consequences belong in the method, even while the tools are internal.
A practice, not a poster
Build the case. Test the conviction. Correct the record.
The principles matter when they shape the method, the published claim and what happens after new evidence arrives.