From the NSA to OpenAI: General Paul Nakasone’s Appointment and the Future of AI Governance
This morning, I noticed that retired 4-star Army General and former NSA Head Paul Nakasone was appointed to the Board of OpenAI. I had an immediate, visceral, and conflicted reaction.
General Nakasone’s credentials for such a role are undeniable. Having headed both Army and US Cyber Command, preeminently countering the massively complex global threatscape at scale, who better to protect the current and potentially leading company in AI? It is the other side of the NSA’s mission and relationship with both the intelligence and military communities that is of note.
It is exactly this complex, ambiguous, and impactful technology that requires General Nakasone’s skilled leadership. OpenAI’s foundational notion is Artificial General Intelligence. It is OpenAI’s stated mission: “Our mission is to ensure that artificial general intelligence—AI systems that are generally smarter than humans—benefits all of humanity.” But it is also trust in the CEO and the ability of the Board to provide oversight, safety, and ensure that OpenAI acts in society’s best interest. Considering how impactful OpenAI is, and how it’s becoming more and more capable each day, this Board holds the keys in many ways to our technological future. Right?
Maybe. As a student of military history, a board member, and a technologist, there is a part of my mind that cannot shake my discomfort. Boards in essence are supposed to be truly representative of the community the company serves. AI has become the next battleground of non-state and irregular warfare, of offensive and defensive actions in cyberspace. In cybersecurity, the “bad guys” are clearly using AI to riff on old tricks, but the truly adept are using it to create new methods. Just as easily as AI can make trivial work of adding hidden information within seemingly innocuous images, deeper meaning and representation can be hidden within the overt action of adding a cyberwarrior to the ranks of the OpenAI Board.
Is this asset far too important to evolve in the hands of a purely civilian board? What civilians know about military and intelligence operations is probably best represented by a drop of water in a giant lake. How can a technology, which clearly has impacts far beyond civilian application, be managed effectively without empowering our state and non-state competitors?
I am both excited and tepid about the appointment of General Nakasone to the Board of OpenAI. He is a highly skilled, complex thinker who has dedicated his life to service. He is exactly the type of leader and person I would want in such a role. It is important that he contribute to OpenAI’s transparency, safety, and security in positive ways. Yet I have significant questions about the direct involvement of someone who has a very clear relationship with military and intelligence circles.
These are challenges I wish I could answer. Sorting through the hype of AI, which is so overdone at points it makes me wince when I hear those two letters, I also feel the same level of impactful societal changes as I felt when I opened up my first web page in 1993. Its potential blew my mind. It evoked both amazement, but also a feeling of “this could be more”. I looked back at that moment and realized how clunky and, by today’s standards, so almost caveman HTML 1.0 looked.
To me at the time, it looked like OpenAI (and generative AI) today. It can do so much, yet it really is just predictive text and data cleaning. It has learned from data that was human-generated, but in the future, those data sources will be (and maybe already are) largely polluted by AI-generated data. Fresh sources of data to learn from will come from our own interactions, and now with Apple’s AI announcement, I have no doubt even more data from the daily randomness of human existence will be part of that dataset.
As we venture forth in the changeover from generative AI to AGI, far more is at stake. So much positive, and conversely, so much negative will come out of it. It is far more than a group of eight or so board members can even comprehend, no matter their brilliance. We cannot fathom the depths of change we are about to experience. And to be clear, that is not hype, but reality. I wish General Nakasone and the rest of the Board of OpenAI the best possible outcomes.
I have hope that it will lead to some of the most amazing benefits the world has ever witnessed. It is crucial that we navigate these challenges with transparency, accountability, and a commitment to the greater good of society. The future of AI holds both immense potential and significant risks, and it is up to those guiding its development to ensure that it benefits all of humanity.
For more information: https://www.theverge.com/2024/6/13/24178079/openai-board-paul-nakasone-nsa-safety
