The Nakada Foundation to Save Humanity exists to stop artificial superintelligence before it ends the human story. Labs will keep building until the law stops them. Law is won in public, not in a research paper.
We are a philanthropic campaign, not a lab and not an ethics board. We do not train models. We do not write papers on chatbot bias, though that work matters elsewhere. Our file has one subject: systems that exceed human intelligence and cannot be corrected or stopped. The line we hold is simple: humanity must never build a superintelligence.
We copy the campaigns that already moved the world when the science was clear and the politics were not: landmines, nuclear non-proliferation, the ozone layer. In each case the data arrived first. Organized pressure finished the job. That is the machine we are building.
We take no money from AI labs, their investors, or firms whose books depend on racing the frontier. We do not staff the work with people whose next job requires those labs' goodwill. Independence is a precondition, not a slogan.
We want narrow AI that diagnoses cancer and speeds drug discovery. The fight is not with "AI" as a category. It is with general systems built to exceed and replace human reasoning across every domain.
The fix is political, not a better loss function. Alignment research helps and still cannot carry the load alone. What works is international law with inspectors and costs for cheating. Nuclear weapons, chemical weapons, and the ozone layer already proved the shape. When Ronald Reagan watched the 1983 film The Day After, the abstract war became a concrete one, and he later signed the treaty that killed off a whole class of missiles.
We sit outside the industry on purpose. The job is a coalition that turns safety from a preference into law, advocacy that moves what counts as serious, and culture that makes the crisis hard to shrug off.
"AI safety" has stopped meaning anything in particular. The phrase covers chatbots that produce slurs, and it covers the chance that a machine smarter than any person ends the species. Those two problems share a label and almost nothing else. A term that broad lets everyone agree while committing no one.
Most foundations name themselves after a method or a field of study. The language is measured and built not to alarm. It works. Nobody who reads it feels any urgency.
"To Save Humanity" states the job. Some of the researchers who built modern AI have put the chance it causes human extinction this century at ten to twenty percent. A risk described in those terms does not deserve careful institutional language.
The name is meant to be impossible to misread. Someone who sees it once knows what the work is for, with no briefing and no glossary.
The Nakada Foundation to Save Humanity was established in 2025 by someone with conviction rather than credentials. That is how most movements start. Abolitionists rarely held history degrees, and suffragists rarely held law degrees. They read the moral stakes correctly and refused to wait for consensus.
The AI safety community has spent a decade building consensus inside academia and inside the industry. That consensus now exists. Geoffrey Hinton, Yoshua Bengio, Stuart Russell, and hundreds of the most respected researchers in the world have concluded that the development of artificial superintelligence poses a genuine risk of human extinction. Their consensus has not produced binding law.
The reason is structural. The researchers who understand the risk work inside or alongside the laboratories building the technology. Their funding, their careers, and their professional relationships are entangled with the very institutions they are warning about. Independence is nearly impossible within that structure.
The Nakada Foundation was founded by Naoto Nakada, an entrepreneur and marketing agency founder who has lived and worked across Tokyo, Sydney, Los Angeles, London, and Singapore. A believer in the philanthropic value of narrow AI, Naoto's concern is with systems built to exceed and replace human intelligence. He brings the instincts of a movement builder, someone who expects the world to change through law and culture rather than papers and conferences.
We must safeguard our collective light of consciousness, a tiny candle in a vast darkness.
No. We support narrow AI: systems that perform specific tasks and can be corrected. What we oppose is artificial general intelligence and artificial superintelligence, whose autonomy and open-ended goals make them categorically different from any previous technology. The smartphone, the internet, even nuclear power can be regulated and shut down. A misaligned superintelligence cannot.
This is the most common objection, and it is precisely the argument that was made against nuclear non-proliferation. The NPT is imperfect, but it has worked. No nation has used nuclear weapons since Hiroshima and Nagasaki. Today, nine countries have nuclear weapons when hundreds could. The answer to "what if others defect?" is a treaty with verification, not a race. The compute required to train frontier AI systems comes from a small number of identifiable chip manufacturers. Monitoring and enforcing a compute threshold is technically feasible. It requires political will, not technological breakthroughs.
Yes. They also have sales teams, investor relations departments, board members who represent shareholders, and competitive pressure to ship products faster than their rivals. Safety teams inside commercial AI labs answer to the same incentives as the rest of the company, and those incentives reward shipping capability over reducing risk. When OpenAI's safety team published concerns about moving too fast, the company moved faster anyway. At best an internal team complements external regulation. It cannot stand in for it.
The Turing Award winner Geoffrey Hinton, who spent his career building the foundations of modern AI and is arguably the most credentialled person alive to make this judgment, estimates a 10 to 20 percent chance that AI causes human extinction within the century. Nobel laureate Yoshua Bengio has stated publicly that the risk is real and underappreciated. The Center for AI Safety's 2023 statement, signed by over 500 AI scientists including Sam Altman, placed AI extinction risk alongside nuclear war and pandemics as a global priority. When the people who built the technology are afraid of it, that is not a drill.
Join the Foundation. Share our message with people who haven't heard it. Contact your elected representative and ask what their position is on AI risk and what legislation they are supporting. If you have financial resources, consider donating to the Foundation or to other organizations working on AI safety policy. If you have cultural or media influence, use it. If you work inside an AI laboratory and have concerns, consider speaking publicly. Political will decides this issue, and political will comes from people: from you, and from everyone you can persuade.
Stay informed. Be heard. Help build the political force this cause requires.