Stefan Leijnen
Utrecht University of Applied Sciences
N2 Conference 2023
University of California, Berkeley
Purposeful AI between Biosphere and Noosphere
Artificial intelligence developed within the machine learning paradigm is capable of producing models of unbounded complexity, far beyond the scope of human understanding. Deep learning, attention networks and foundation models have pushed subsequent shifts in this capacity, scaling up model size, data usage and computing power to a point where this complexity has become both a bug and a feature. This has driven human developers and users towards progressively relying more on blind trust, humility towards perceptions of self-autonomy and problems of control.
A common assumption in the machine learning paradigm is that the objective function (the purpose towards which the model optimises) is provided by a human, securing some level of human control of the AI. However, purpose may emerge not only at the level of human individuals but also in collective, mechanical or hybrid organisations.
AI brings radical innovations to our affordances for interacting with large datasets, such as ChatGPT vs. prompt-based internet search. Interfaces will continue to co-evolve with our use of this technology, exploring and transforming human-machine integrated thought and consciousness through existing and new (e.g. neural) interfaces.
In order to help guide these innovations, we need to develop a framework for understanding the emergence and regulation of purpose in living, artificial and hybrid organisations, i.e. in and between biosphere and noosphere. Using this framework, technology can be designed that allows humans to retain and attain the reigns of control as we continue to delegate autonomy and purpose to machines and other organisations.
Learn more about Human Energy and the N2 Conference 2024 on our website.