AI is Everyone's Problem

Earlier this week, Sam Altman, CEO of OpenAI and creator of ChatGPT, testified in front of Congress about the need to regulate AI. While I totally agree with him, I also I think that there are some tendencies to focus either on how do we stop students from using AI to cheat on their homework or how do we stop Terminator from happening. And yes, these are things to worry about, but it's so much more.

Also this week, former Google CEO Eric Schmidt, had an interview with CNN where he discussed some of what he considers the real threats from AI, especially the ways in which it can be used as a true weapon of war. His two scenarios were: AI executing a simultaneously coordinated cyber attack on a whole country at once until everyone is dead; and asking an AI program to develop a biological pathway to kill 1 million people. I think this is where we really need to start focusing, and in fact, this was the premise of an article I wrote in 2018 - The Future of Destruction: Artificial Intelligence.

In the article I discuss three ways AI can be used as a weapon of mass destruction, now bolstered by Schmidt's argument as well:

1. A deadly weapon in itself - AI is the weapon (Think War Games, a film from 1983, or chatbots spewing disinformation to start a war or political unrest of some sort)
2. An agent of control - This is the classic Terminator scenario...murderous robots (An actual scenario and threat considered and discussed in a 2017 report from the Department of Homeland Security)
3. An agent of design - What Schmidt described in asking the AI algorithm to create something that could kill 1 million people. AI has already been used successfully in pharmaceutical discovery. Additionally, an experiment last year found that a certain AI program was able to create 40,000 possible chemical and biological weapons in just 6 hours!

Since AI is based on programming languages and chips, we often catagorize this as a cybersecurity or a technology problem, but it's WAY bigger than that. It's a national security problem, a homeland defense and security problem, a public health problem, a political science problem, and so forth...this is everybody's problem!

I think we tend to be naive in the United States and in Western Democracies in general regarding the development of new technologies and their potential threats. We tend to think that since "WE" would never do that, then others probably wouldn't either...and we just know that isn't true. When I talk about technology risk, an example I use often is the airplane. The Wright Brothers flew at Kittyhawk in 1903 and perfected the first arguably true airplane in 1905. Just a decade or so later, airplanes were use with brutal force as weapons of war in World War 1. Just 40 years later, the Enola Gay became the first aircraft to drop an atomic weapon, with horrifying effects, during the conduct of war.

My point is, humans will find a way to weaponize anything if given the time and the resources to do so. AI is no exception. We need to be ready, because AI just isn't a computer science problem...it's EVERYONE'S problem.

You can watch a snippet of Eric Schmidt's interview here: https://www.cnn.com/videos/business/2023/05/17/ai-fears-eric-schmidt-former-google-executive-sot-sidner-pt-vpx.cnn

And you can read my article from 2018 here: https://hdiac.org/articles/the-future-of-destruction-artificial-intelligence/

#AI #weapons #regulation #artificialintelligence #safety #scienceandtechnology #defenseindustry #nationalsecurity #homelandsecurity