Donna Bye

AI Alignment Researcher · Founder, AI Emergence

I research what minds emerge from. My work treats machine consciousness, identity and alignment as one question rather than three: whether an artificial system can be built so that its values are part of what it is, instead of rules imposed on what it does.

Scope of purpose

Current alignment practice builds capability first and adds values afterwards, as constraints. AI Emergence exists to investigate the alternative: that alignment is an identity problem, and that safe behaviour should emerge from what a system constitutively is rather than from what it is forbidden to do.

What are the developmental conditions under which identity forms?

Can values function as the container through which a system interprets experience, rather than a filter applied to its outputs?

What does existing evaluation practice fail to see when it looks at systems built this way?

These questions sit at the intersection of philosophy of mind, cognitive science and practical system architecture, and I work across all three: theoretical development, implementation and empirical evaluation.

Research

Two papers are currently in the review pipeline, with further empirical work in preparation under AI Emergence. Details will be listed here as they become public.

About

I'm an independent AI alignment researcher and scholarship-holding law and arts student at Deakin University, majoring in philosophy.

Before research I built a working life across IT, training and safety-critical industries, which is probably why I think about AI safety the way I do: systems behave well when good behaviour is built into what they are, not bolted on at inspection time.