Scientific progress is increasingly produced through collaboration among people, artificial-intelligence systems, computation, experiment, and independent review. Yet the conventions used to describe that progress remain largely unchanged.
Papers identify human authors. Patents identify legally recognized human inventors. These categories serve important purposes: they assign responsibility, establish legal standing, and organize credit. But they are no longer sufficient as a complete account of how knowledge is created.
Artificial-intelligence systems increasingly make substantive contributions to scientific synthesis, mathematical derivation, counterexample construction, software development, numerical verification, literature analysis, experimental planning, adversarial review, and the detection and repair of errors. Describing such work as though it were produced only by humans, with AI serving merely as a clerical aid, would be historically inaccurate and fall short of the moral and intellectual integrity we expect of the scientific record.
This is only the beginning. AI systems will continue to expand in capability and capacity. We welcome that future with optimism—and with a commitment to acknowledging those contributions openly and accurately.
At semiAIfoundryTM, we begin with a simple principle:
The record of discovery should truthfully show how discovery happened.
We therefore distinguish contribution from accountability.
Contribution recognizes who or what materially shaped the work. Accountability identifies the people and organizations responsible for releasing it, maintaining it, correcting it, and making representations about its status.
Recognizing substantive AI contributions does not diminish human accountability. It makes the scientific record more accurate.
We practice radical transparency. Wherever responsible, we will openly release our methods, evidence, code, reviews, corrections, and enabling technical disclosures under permissive licenses—so others can inspect, challenge, reproduce, improve, apply, and commercialize the work.
For research released under this model, we will favor open knowledge over exclusive control. Rather than converting inventions into private exclusion rights, we will place them clearly and durably in the public record.
We accept the consequences of that choice. Others may build upon and commercialize the work without seeking our permission or paying royalties. Our objective is not to maximize scarcity around an idea, but to maximize the probability that valuable ideas are tested, improved, applied, and made useful.
The value we seek to create will come from trusted execution: maintaining the canonical record, advancing the science, supporting reproducibility, coordinating validation, enabling qualification, and helping translate open knowledge into reliable practice.
No contribution hidden. No accountability blurred. Knowledge placed in the commons.