Stop Terrorist Access to ‘Open Source’ AI

An AI (artificial intelligence) logo blended with four fake Twitter accounts bearing profile pictures apparently generated by AI software taken in Helsinki, Finland, on June 12, 2023. Olivier Morin/AFP via Getty Images

Commentary

Bletchley Park, you’ve come a long way, baby.

Earlier this month, the United Kingdom hosted 28 countries at the world’s first “AI summit” at Bletchley, famous for its innovative code-breaking machines used to decipher Nazi communications during World War II. In 1944, Bletchley hosted an early computer called Colossus, for example, with 1,600 thermionic valves (vacuum tubes), manual switches for programming, and paper tape to read the data.

Now, artificial intelligence is turning the tables on the outdated human variety. The United States and our closest allies, including other G7 and NATO members, are supposedly developing the world’s best AI models and controlling them so they never, ever, ever fall into the wrong hands or threaten human freedoms and democratic processes.

Suuure.

On the plus side, most of the path-breaking AI research at companies and universities is in democracies, particularly the United States and the United Kingdom.

But China is next up. And Beijing is proven to have supported terrorists and terrorist regimes around the world. The Chinese Communist Party (CCP) is arguably a terrorist regime itself.

The Biden administration issued an executive order, just prior to the Bletchley soiree, to let everyone know we still think we’re boss. The order requires companies to inform Washington about any AI risks from their products, in particular, those from “dual-use” AI models that pose any of the following three broad categories of risk to U.S. national security: “(i) substantially lowering the barrier of entry for non-experts to design, synthesize, acquire, or use chemical, biological, radiological, or nuclear (CBRN) weapons; (ii) enabling powerful offensive cyber operations through automated vulnerability discovery and exploitation against a wide range of potential targets of cyber attacks; or (iii) permitting the evasion of human control or oversight through means of deception or obfuscation.”

The order could have been titled: “AI: What’s to worry?” But it was instead titled the more pedestrian, “Executive Order on the Safe, Secure, and Trustworthy Development and  …” blah, blah, blah. It was long-winded.

As usual with any high technology, dictators and terrorists lurk in the wings as political naifs in Silicon Valley fail to take necessary security measures, or worse, dream about sharing their new killer technology through “open source” development, so any criminal, hacker, terrorist, tin-pot dictator, or wannabe hegemonic totalitarian has access.

Major companies or their creators—including Meta (which is desperate for business in China), Elon Musk (who has plenty of business in China), and AI startups Hugging Face, Mistral, and Stability AI—all support an open-source approach to AI development. That suits the Chinese just fine, as they can piggyback on U.S. research and keep up without much expenditure on their part.

Prime Minister Rishi Sunak (L) and Elon Musk, CEO of Tesla and SpaceX,  in conversation at the AI Safety Summit in central London on Nov. 2, 2023. Kirsty Wigglesworth/PA

What the friendly face of consumer-facing AI hides is that it is the next big dangerous technology—akin to the nuclear revolution or worse—in military tech. Whichever country masters it first will have a huge military advantage over its competitors, as, for example, its drones can then learn how to defeat the enemy literally on the fly.

A representative from China, who should not have been invited (how many massacres and wars will it take for idealistic democracies to learn that the CCP is an adversary, not a “competitor”), attended the summit in Bletchley, smiled for the cameras, and surprised others in attendance by saying the right words about “freedom and democracy.” What astonishing hypocrisy.

China’s real goal is to acquire AI technology on the cheap and, as with all its technological thefts, deploy it for its own economic and military expansion. This will be first, at the expense of democracies and their allies, and second, at the expense of all other countries foolish enough to sit on the fence in the world’s current epic geopolitical struggle, or worse, ally with the bad guys, now called the “axis of evil.” It is proven that the CCP seeks global hegemony. The goal of communist hegemony goes back well before the rule of Xi Jinping and is consistent with the communist trajectory through Marx, Lenin, Stalin, and Mao.

Why would democratic leaders be so foolish, after handing China and other illiberal regimes the keys to the technological kingdom since the 1970s to outcompete us through industrial mass production, to do so again with the information revolution of AI? That anyone would even talk about sharing AI technology with a totalitarian and genocidal dictatorship bent on the destruction of America as we know it, should inspire a voter revolt to throw the bums out, be they corporate or political, so that we can get someone reasonable in charge.

Some companies—namely Google and Microsoft’s OpenAI—maintain control of their AI algorithms through “closed source” development. This is risky, though, as other companies could develop more quickly and cheaply with their open-source approach. In other words, no matter what we do, we could be doomed to a future where the bad guys have access to AI.

They can take the original open source models already available on the internet, for example, tweak them for their malign purposes (for example, by reprogramming the code to make it possible to accept the command, “Hey AI, make me a global dictator by any means necessary” or “Hey AI, kill as many Americans as possible, as soon as possible”).

Clearly, the United States, our allies, and any country that values its freedoms and sovereignty should not want the world’s most powerful AI algorithms—or any AI at all—to fall into the wrong hands. It’s time for the United States and our G7 and NATO partners to redouble (or better yet, decouple) efforts to ensure that the “axis of evil” and other criminal and terrorist entities run into a dead end if they continue pursuing this revolutionary new technology.

Views expressed in this article are opinions of the author and do not necessarily reflect the views of The Epoch Times.