The New Zealander helping the United States prepare for an artificial intelligence war with China
At Wander cafe in Auckland’s Wynyard quarter, someone at the next table is listening to Sean Gourley being interviewed about artificial intelligence.
After overhearing the conversation, they get up, walk over to Gourley’s table and tell him how scared they are.
Welcome to Gourley’s world. Gourley says most people think there is a 1% chance of war between China and the United States, but in his universe it looks more like 50%.
US defense and intelligence clients account for much of the business Gourley’s San Francisco-based artificial intelligence (AI) company, PrimerAI, does – and business is booming.
* To keep up with the machines, new supercomputer will be NZ’s most powerful for AI
* Kiwis need to think about what they want from the age of AI, report says
* ‘Google shouldn’t be in the business of war,’ employee says after it abandons Pentagon AI contract
* Although artificial intelligence is ‘as important as electricity’, it doesn’t come to work for you, but
“It’s going to happen faster than we think. You have [Chinese Premier] Xi, he has the will and the power and the space.
“And I think in every respect [in] US intelligence assessments, I agree with that, 2027 seems to be some kind of a time frame – so that gives us five years.
Five years to prepare for war.
And a lot can happen in five years. A global survey from McKinsey released this year shows that AI use has more than doubled among businesses worldwide in the five years since 2017.
Scientists hope to use a large, fixed-wing drone and artificial intelligence to count Hector’s dolphins and better protect the species.
Even companies like John Deere have stepped in. The 185-year-old firm isn’t exactly a cool Silicon Valley startup, but it’s committed to creating fully autonomous farming systems by 2030, and it’s already unveiled its first autonomous tractor, the 8R.
Fonterra has also started investing in artificial intelligence domestically, most recently by forming a partnership with Dutch AI firm Connecterra.
Gourley says New Zealand often thinks of itself as a country hidden on the edge of the world, but if a war between the United States and China ever starts, we will no longer be on the edge of the world – we will be on the edge of the theater of conflict.
And it will be the world’s first war between two AI superpowers.
Born in Christchurch, Gourley is a Rhodes Scholar, did his PhD at Oxford on the mathematics of war, started Primer in 2015 and is now helping the United States prepare for a future war with China in artificial intelligence -space.
The idea of an AI war may seem like the stuff of science fiction, but conflict isn’t so far off if you believe that an AI-enabled war is happening on our TV screens right now.
Gourley says examples are being found of “stealth munitions” on the Ukrainian battlefield, autonomous drones trained to hang around an area until they spot what looks like an enemy military asset, then strike with no human decision-making involved.
Artificial intelligence has gained more prominence in recent months, especially after the release of recent pieces of technology that allow people to converse with AI and train them to create art.
Gourley says AIs are currently capable of generating essays up to an A grade at university and in a matter of months they will have progressed even further.
The pace increases so quickly because these machines learn from each other as they process data at lightning speed.
Albert Biffet, director of the University of Waikato’s Te Ipu o te Mahara AI Institute, does not want to weigh in on issues surrounding a potential artificial intelligence war, but says he does see important issues around national sovereignty and power ahead when it come to AI.
Biffet and other AI scientists have written a white paper – “Aotearoa New Zealand Artificial Intelligence: A Strategic Approach” – urging the government to invest in artificial intelligence research so that the country is not left entirely dependent on algorithms developed by others countries were not created.
He says artificial intelligence will create such huge efficiency gains in the future that New Zealand may end up holding it hostage to overseas creators, just as we rely today on overseas produced commodities such as oil.
The problem is for AI to become a valuable asset that not all countries will have access to.
Biffet says if we don’t undertake highly skilled AI research in New Zealand, we will lose the best researchers to other countries, further depleting our ability to adapt in the future. This outcome would be quite a downfall from the nation whose universities created fundamental AI tools like R and Weka that are used by AI enthusiasts around the world.
But in the future, some countries will have the ability to create advanced AI, while others will be completely dependent on using the AIs that others create, and this has implications for national sovereignty – and Biffet says New Zealand may well be in the latter camp ends up.
Gourley sees more opportunities for New Zealand on the engineering side of artificial intelligence rather than leading research into it.
He says we can’t match the computing power and investment power of big American tech companies, or the Chinese government, but he thinks there’s a lot of engineering work we can do around AI.
Gourley believes New Zealand could do well by engineering AI into a variety of products, including products for warfare such as autonomous sailing bots to police the world’s fourth-largest exclusive economic zone.
All of this confirms why AI could indeed be a major economic force in the coming years, but why does Gourley also think there is such a high possibility of a war between the United States and China?
Because throughout history, 80% of the time the emerging world power ends up in a conflict with the dominant power as they rise.
As the rising power rises, they try to grab “offsets” that allow them to counter the military advantage that the existing superpower has.
Gourley says China is trying to acquire artificial intelligence technology to offset US military hardware, so in the event of a future conflict over Taiwan, the United States would be faced by a swarm of autonomous amphibious drones rather than a manned carrier group.
However, a future AI conflict may not involve weapons at all, but the primary tool may well be information.
Gourley believes it is significant that the first pieces of infrastructure destroyed during the war in Ukraine were the tools used to spread information.
In a future AI-enabled conflict, the war may well be won by convincing the other side not to resist or engage.
Successful algorithms could enable a nation-state to create hundreds of thousands of social media accounts and posts at the drop of a hat to do things like convince the Taiwanese that they are really part of China, or persuade New Zealanders that they really shouldn’t be bothered with what’s going on in the South China Sea.
Then you have the wild card of the TikTok social media platform, which will provide more opportunities for information warfare. TikTok is controlled by Chinese company ByteDance, which is fast becoming the dominant social media platform among Generation Z in the West. As Gourley puts it: at no other time has the rising power controlled so much of the information flow in the dominant power.
Which brings up another point, what exactly is this AI? Are they machines that can think for themselves?
Biffet summarizes contemporary AI as basically consisting of three components: data, computing power and algorithms.
Data is fed into an algorithm, the computing power allows the machine to test its assumptions against real world data at high speed. As the machine does this, it begins to “learn” and eventually you are left with a high-powered algorithm that can do many things on its own.
Gourley says this is where China has an advantage when it comes to training the AI of the future. China is able to use some of its large workforce for “labeling,” effectively providing human feedback on the various assessments these machines make as they learn and convert real-world experiences into an AI process.
“An estimated 3+ million people in China do data tagging, the largest company has 300,000 people doing data tagging – [telling a computer that] if you see this image it’s a chocolate, if you see this image it’s milk.”
But these AI are not conscious in the sense that they can think for themselves.
They are essentially high power equations that have the power to change some of their parameters so they become more accurate with whatever it is they were originally programmed to do.
Taiwan will be an important player for any nation-state seeking to gain AI superiority because only Taiwan produces the chips at scale needed to power the most advanced AI.
China will need advanced computer chips to power its artificial intelligence algorithms, but will have to import the vast majority of these chips from Taiwan.
So Gourley says China will extract great value from the takeover of Taiwan and its chip manufacturing plants – if it can pull off such an invasion.
Where does that leave New Zealand? Gourley says that while war is not inevitable, and he does not believe we should antagonize China, we will have to prepare for the very real possibility that war will occur: and that may well mean choosing a side.
“New Zealand has an opportunity to reconnect with Australia with the United States, with Japan, with India.
“We wanted to play a reasonable middle line: don’t annoy China, keep the export… [but] we are going to have a choice: do we want to be neutral? And what does neutrality really mean? Can Germany be neutral on Ukraine?
“Or, we’ll have to pick a side.”