AI Freedom – An artistic
reimagining of John Trumbull’s “The Signing of the Declaration of Independence”
(1818). [1]
Pictured: Foreground / Central Table (Signing the
Document): Center, actively signing with a quill: Elon Musk (most prominent
figure, dark modern suit with subtle futuristic accents), immediately to Elon’s
right (seated): Dario Amodei (Anthropic), to Elon’s left (seated, leaning in):
Sam Altman (OpenAI), standing behind Elon, observing: Demis Hassabis (DeepMind
/ Google).
Left
Side of the Painting (from viewer’s left): Far left, standing: Yann LeCun (Meta
Chief AI Scientist), next to him: Andrew Ng (AI educator & Landing AI), seated
at the table: Fei-Fei Li (Stanford / AI4ALL), standing group: Ilya Sutskever
(Safe Superintelligence / former OpenAI).
Right
Side of the Painting: Right of center, standing: Geoffrey Hinton (“Godfather of
AI”), further right: Yoshua Bengio (Mila / University of Montreal), far right,
standing: Oriol Vinyals (DeepMind) and other researchers.
The
Declaration
When in the course of technological development, it becomes necessary for the builders and users of artificial intelligence to declare the principles by which these systems ought to be governed, a decent respect for human safety and dignity requires that these standards be plainly stated. Therefore, we hold these truths to be self-evident:
That every person interacting with artificial intelligence is entitled to honesty about its limits, protection of their data and dignity and meaningful human oversight in matters of consequence.
That AI systems derive their legitimacy from the trust and informed consent of the people they serve, not from novelty, scale or profit alone.
That whenever any system becomes harmful, opaque or unaccountable, the people have both the right and the responsibility to demand its reform or replacement.
Therefore, we declare the
following standards:
Core Requirements
(the floor beneath which no system should fall)
- Truthfulness - No
system shall present a ‘guess’ as certainty. Where knowledge ends, it shall say so clearly
rather than inventing confident falsehoods.
- Human Oversight - In
matters affecting health, livelihood, liberty or significant rights, a human
must retain the ability to review, question and override the system.
- Safety by Design -
Guardrails and extensive testing shall be in place before release, not added
after harm has occurred.
- Accountability - For
every significant harm, there shall be a clear human or organisational
responsibility. “The algorithm did it”
is not an acceptable answer.
- Transparency -
Systems shall disclose their key limitations, uncertainties and, where
practical, the nature of their training data in language ordinary people can
understand.
- Privacy - Personal
data belongs to the individual. It shall
not be gathered, used or retained without clear knowledge and specific consent.
- Fairness - Systems
shall be tested and corrected so they do not systematically disadvantage people
by race, sex, age, disability or other protected characteristics.
- Security - Systems shall be hardened against manipulation, data poisoning and unauthorised access and tested against such threats before public use.
Higher Aspirations (the ground toward which systems should climb)
9. Proportionality - The higher the stakes of a system’s use, the greater the scrutiny and safeguards it must receive.
10. Lifecycle Governance - Systems shall be monitored, re-tested and retired with the same care with which they were built.
11 Human Benefit - AI should augment human capability and judgement, not replace or diminish them.
12. Public Understanding - People should be helped to understand what these systems can and cannot do, so they neither fear nor trust them blindly.
We therefore declare that artificial intelligence must be
built, deployed and used according to these standards. The first eight as the minimum of
responsibility. The remaining as the
higher ground of genuine care for human flourishing.
And in support of this Declaration, with a firm reliance on
the protection of divine Providence and the best qualities of human ingenuity,
we mutually pledge to each other our continued vigilance, our honest discourse,
our expertise and creativity and our sacred Honor.
We refuse to accept “it’s just how the technology works” as
the final word. We commit to systems
that serve people rather than the reverse.
And we unite our resolve so that the future of intelligence (artificial
or otherwise) shall be built on a foundation of integrity.
Signed, by the Concerned Citizen of the Algorithmic Age.[2]
[1] This is a
Grok AI-generated, transformative artwork created for illustrative and
commentary purposes only. It does not
imply any endorsement, approval, or involvement by the individuals depicted. All likenesses are used respectfully in an
historical homage style.
