AI in National Security Conference
Day 1 quick take
Washington DC 7 May 2024
There is a well attended AI conference in DC today and tomorrow. It has a strong NatSec bent. Multiple simultaneous panels, so there is no way anyone can see it all.
Here is MIL’s quick take on what he saw and heard.
Dr. Liz Sherwood-Randall raised the issue of CRISPR + AI. If COVID scared you, if CRISPER concerns you, and if you know enough about AI to be dangerous; this topic will turn you into a recluse washing your hands every ten seconds (not that it will save you this time). Dr S-R has gravitas and, while highly scripted, did a good job highlighting key issues and outlining the Administrations policies (as one might expect).
However, I could not help but feel in a parallel universe where serious people talk policy about serious threats to national security - in stark contrast to everything else in this city, especially around Capitol Hill, where Senators auditioning to be Pence 2.0 are already rejecting the outcome of the next election - openly and blatantly promising Coup 2.0.
Its was bizarre and jarring to see “normie” policy heavyweights speak about real issues, knowing Congress has an important role to play and half of the House and Senate are planning to continue to do the exact opposite of responsible governance.
The "world by 2050" Panel was a disaster. Listening to the experts on stage, everything will be linear and predictable. Any serious issues will be external to the US. It felt like listening to a conference in 1938 where no one wanted to address the crisis in Europe and instead talked about vague projections about the 1950s where concerns were limited to policy alignment challenges - interest rates or inflation a bit too high or low, or migration being more of a challenge, or climate change being a tad annoying. When the moderator Bina Venkataraman (Washington Post) pressed the panel about AI being used for Dis/Mis/Mal information and manipulating western populations and turning them further against democracy the panelists wandered off to talk about something anodyne. I tuned out.
What should they have said? Simple. The Western world stands on a precipice of a new dark age where authoritarianism has been waging a war against the great democracies for over a decade. They are winning as evidenced by the war-blindness of the experts charged with thinking about the future presented on this panel. Adding the incredible powers of AI to Mal/Dis/Mis-information in unchecked social media systems empowers the autocrats at home and abroad to win. In the context of the unfolding Coup 2.0 this challenge only grows by the hour. The groupthink of this panel had no clue that Coup 1.0 had happened or what it meant for global security. It was a stunning failure of serious people in positions of responsibility.
The Future of Compute for AI Advancement, largely inaudible, promised to discuss “transformative potential of generative AI and how it shapes the trajectory of the coming decade. From breakthroughs in chip design to the implementation strategies of real-time AI infrastructure, they will delve into the evolving landscape of generative AI and Large Language Models to explore the profound impacts on supply chain security, national defense, workforce productivity, and global climate challenges.” and instead spent all their time talking about the energy consumption patterns of data centers. Important but will be managed in time and not what was promised. Fail.
Transforming Decision Making in the IC was a fun quick hit with Jonathan Ross. He basically had a chat with an AI about what she could do to enhance data collection, sorting and analysis. The interaction was well managed, it had some glitches, but Ross powered through and stayed on theme. It felt like exposure to a closed source AI - higher quality than an open AI - and that alone was interesting. The AI was quite conversational and responsive but after the novelty wore off it was just a fast verbal rendition of any other AI’s canned responses. It usefully showed the pluses and minuses of AI
By far the most entertaining and free wheeling panel was The Future of Geopolitics and the Role of Innovation and Technology. It turned into a showcase for Alex Karp to “speak freely” which MIL found refreshing. His core point was he was worried about America hiding from being #1 in AI development, this was a significant advantage we should not be scared to be proud of - or use. He noted the US still leads AI development by a substantial margin, likening it to the development of Jazz in the 1940s and 50s. Sure Germany or Japan might have a few bands, but the center of gravity of Jazz was and remains the US.
His bigger bugbear is what a friend of MIL called “Hamas chic” particularly on American campuses. He sees that as creating a double standard regarding how America and its enemies are judged. In other words, America must behave perfectly while Hamas can cut peoples heads off and that’s acceptable behavior.1 What MIL found particularly interesting about Karp’s carping, was he was blind (or at least neglected to include the critical issue of) the enemy’s use of social media and AI that created this double standard in the first place.
General Milley outlined his perfect solid character of war narrative. Namely that the character changes but the nature does not. Agreed. Pure Clausewitz. End of discussion. He rightly mentioned using the blunt tools of violence to force others to do our will.
What the panel missed… is the motherlode of AI with respect to war. And its not robots!
What if the other means of politics is NOT to force the enemy to do one’s Will. What if AI offers a pathway to convince the enemy to adopt one’s narrative as their own?2 Thats the strategic game changer on offer in AI.
This is a subject of a forthcoming MIL article that has been in the works for some time.
MIL was eager to see the Beyond the Hype panel but some kind of crisis had everyone heading for the doors. MIL was well on his way out of the zone when the unexpected all clear came back.
As they say, tomorrow is another day.
America has always struggled with realism and idealism in its foreign policy.
This has already happened - Putin has turned the party of “tear down this wall” Ronald Reagan into a pro-Kremlin cheer squad living in the land of upside down where Ukrainians are Nazis who must be crushed by America via cutting off all support.