The tools of global spycraft have changed
And so has the world in which they are used
A FEW years ago, intelligence analysts observed that Internet-connected closed-circuit television cameras in Taiwan and South Korea were inexplicably talking to vital parts of the Indian power grid. The strange connection turned out to be a deliberately circuitous route by which Chinese spies were communicating with malware they had previously buried deep inside crucial parts of the Indian grid (presumably to enable future sabotage). The analysts spotted it because they were scanning the Internet to look for “command and control” (C2) nodes – such as cameras – that hackers use as stepping stones to their victims.
The attack was not revealed by an Indian or Western intelligence agency, but by Recorded Future, a firm in Somerville, Massachusetts. Christopher Ahlberg, its boss, claims the company has knowledge of more C2 nodes than anyone in the world. “We use that to bust Chinese and Russian intel operations constantly.”
It also has billions of stolen log-in details found on the dark web (a hard-to-access part of the Internet) and collects millions of images daily. “We know every UK company, every Chinese company, every Indian company,” says Ahlberg. Recorded Future has 1,700 clients in 75 countries, including 47 governments.
The Chinese intrusion and its discovery were a microcosm of modern intelligence. The Internet, and devices connected to it, is everywhere, offering opportunities galore for surveillance, entrapment and covert operations. The entities monitoring it, and acting on it, are often private firms, not government agencies.
What is made possible online, though, does not stay online. The cameras in Taiwan and South Korea were among more than one billion installed globally, part of a mushrooming network of technical surveillance that has made life harder for intelligence officers and the agents they must recruit.
Intelligence services spend their time obtaining the secrets of others, protecting secrets of their own, and engaging in covert activities that other arms of government would rather disclaim: building influence, maintaining secret contacts with terrorist groups, disrupting plots or assassinating foes. Technology has been integral to this for more than a century.
As the world shifted from couriers to telegraph to radio, writes Michael Warner, a former Central Intelligence Agency (CIA) historian, intelligence became “mechanised and industrialised”. It also became a consumer of the highest tech available.
The first programmable computer, Colossus, was developed to crack Nazi codes. The advent of spy satellites laid bare Earth’s surface. Bleeding-edge gadgetry, such as miniature transistors and secret ink derived from the emerging plastics industry, changed how human intelligence was done. All these trends were most notable in the world’s great technological superpower: America.
Then, as now, some of that technology came from outside the spies’ world. When the CIA held one of its first artificial intelligence (AI) symposiums in 1983 at Langley, Virginia, it included a vendor show where industry could show off products. But the flagship technologies – supercomputers, spy planes, the Corona satellites – were elite projects, mostly used to spy on a world that did not have that technology, and inherently statist, by dint of their scale and expense.
What is different today is that digital technology has seeped into every corner of life. There are now 8.5 billion smartphones worldwide, part of a microelectronics revolution in which the cost of cameras, microphones, GPS receivers and other sensors has plummeted.
As global Internet penetration soared from 35 per cent in 2013 to 67 per cent in 2023, these sensors – in phones, computers, cars and televisions – were knitted together, creating intelligence factories in every pocket, on every street. The algorithms and computing power to make sense of the torrent of data grew exponentially more capable and efficient.
That has changed the things that spies care about. Intelligence services are tasked not only with keeping tabs on the technology of their rivals – is a rival building a large AI model in secret? – but also with shaping how it is used.
Western security services fret that Chinese influence over products connected to the Internet, from TikTok to routers to self-driving cars, could become tools for data collection, political meddling or wartime sabotage. Signals-intelligence (Sigint) agencies were at the forefront of public debates over whether Huawei, a Chinese telecommunications giant, should be barred from supplying key parts of Western infrastructure.
As well as changing what is there to be spied on, it also changes how spies work. Some things are easier. In the Cold War, covert operations – or “active measures” in KGB parlance – required elaborate and expensive infrastructure, notes Thomas Rid of Johns Hopkins University.
The CIA’s paramilitary operations required deniable military hardware (old Soviet guns), front companies (fake airlines) and a willingness to risk American lives. Virtual operations – such as Russia’s recent targeting of industrial control systems in American and European power and water networks – are cheaper, easier and safer.
“Computer-network operations have a built-in temptation to make covert action more destructive, more aggressive, and to veer from subversion to sabotage,” says Rid.
Now you see me
Other operations have become harder. The spread of end-to-end encryption has complicated Sigint. The growth of “ubiquitous technical surveillance” (UTS) has made it harder to conceal the identities of intelligence officers, harder for them to travel under alias and harder for them to meet agents without getting caught. That scrutiny is more potent, more widespread, and more hostile: China has exported surveillance technology worldwide.
These are not merely arcane games of spy versus spy. Hamas’ invasion of Israel on Oct 7 laid bare the consequences of intelligence failure. In a world where great-power war looms larger, good intelligence can help bring stability. In the Cold War, the North Atlantic Treaty Organization wanted Warsaw Pact Sigint to be “decent or good”, notes John Ferris in his authorised history of GCHQ, Britain’s Sigint agency, because it helped deterrence.
This Technology Quarterly considers how UTS, the data deluge and advances in AI are touching every branch of spycraft, from the geospatial analysts who must wade through an ever-growing pile of satellite images to the Sigint agencies who must eavesdrop on a largely encrypted world. But it begins with cloaks, daggers and the humans who wield them.
©2024 The Economist Newspaper Limited. All rights reserved
