Safety
The two faces of Janus
What we ask for must match exactly what we deserve, with no undesirable add-ons embedded maliciously.
Satellite tech and AI safety
Projects
An open evaluation of whether AI safety holds up in Nigerian Pidgin. Results due November 2026.
My vision · three pillars of AI safety
I am Ishaku. I care about satellite technology and AI. I want to see a world without disease, with abundant healthy food, and with a strong relationship of trust, safety and security between AI and humans.
Three pillars, each told through a story far older than computers.
The two faces of Janus
What we ask for must match exactly what we deserve, with no undesirable add-ons embedded maliciously.
The night shift
When no one is watching, will AI remain trustworthy?
In Plato’s Republic, the Ring of Gyges makes its wearer invisible, and asks whether anyone stays good when unseen. A trustworthy AI does the same careful work in the dark as under the brightest lights.
Ubuntu: I am because we are
AI should warm the whole circle, and never break it.
The Nguni saying umuntu ngumuntu ngabantu means a person is a person through other people. Deploy AI like the fire at the centre of the circle: shared and tended together, pluralistic and community-centred, never robotic or profit-first.
What it's for
Why safety matters at all.
Faster diagnosis and discovery, reaching every clinic.
Smarter farming, better nutrition and early warnings, from AI and satellites.
AI that raises productivity and incomes, and spreads the gains widely.
What stands in the way
Safety fails both ways: helping the wrong people, refusing the right ones.
Too little African data means models that understand Africans poorly.
Safeguards that hold against the same attacks that keep coming back.
Ramblings
About & contact
I’m Ishaku, a satellite tech engineer moving into AI safety. I’m building Salecta, an open test of whether AI safety holds up in Nigerian Pidgin, and I’m looking for a fellowship or role where I can do this work full-time.