“The road to AGI should be a giant power struggle.” Sam Altman (2024), CEO and Board Member of OpenAI

…

In September 2023 I returned to Durham University and enrolled in the Business School’s Management (MSc) programme. The year following has been one filled with revelation and great contentment. It has been a time filled with extensive learning and consolidation, both on the programme and off of it, and a time I am deeply grateful for.

The Management programme itself was the most challenging experience of higher education. The first two terms were taught, covering many key facets of business management (including strategy, accounting and economics, strategic marketing, organisational behaviour, and supply chains) whilst the third term and summer were spent researching and writing a 15,000-word dissertation.

Owing to my keen engagement with issues related to AI (and AI safety), I decided early on that I wanted to centre my dissertation on this field. AT present, corporations are developing the world’s most powerful AI models and returning to the opening quotation above, Altman’s prophesising may well be validated. Who this power resides in the hands of and what they choose to do with it will have profound impacts on all aspects of human life.

From a management perspective, the novel and untested corporate governance structures debuted by multi-billion dollar start-ups OpenAI and Anthropic rang a number of alarm bells of interest. Further reading of the academic literatures on AI safety, corporate governance and legitimation revealed a chasm at their intersection, devoid of any meaningful academic contribution.

I am of the belief that to safeguard the future of humanity we must build safe AI systems, and to do so we must know where power lies and how it is utilised by varying actors to legitimate themselves. The research that was conducted here sought to provide the first batch of caulk with which to seal this academic void.

The publication below represents hundreds of hours of research and writing. The process was without doubt the most intellectually challenging task I’ve ever attempted, but equally without doubt also the most rewarding.

If you do give it a (skim) read, then I thank you sincerely for taking the time to do so.

Wishing you a good day, or a good night, and a good life.