Why AI Makes Computer Worms Harder to Stop

Letters spelling "AI," a keyboard and a robotic hand in this illustration created on September 23, 2026. (Reuters)
Letters spelling "AI," a keyboard and a robotic hand in this illustration created on September 23, 2026. (Reuters)
TT

Why AI Makes Computer Worms Harder to Stop

Letters spelling "AI," a keyboard and a robotic hand in this illustration created on September 23, 2026. (Reuters)
Letters spelling "AI," a keyboard and a robotic hand in this illustration created on September 23, 2026. (Reuters)

Anthropic's warning that AI-powered "worms" could soon spread across smart gadgets is the latest in a series of studies showing how artificial intelligence could supercharge self-spreading malware.

A computer worm is malicious software that spreads between devices on its own.

In June, Nicolas Papernot, an associate professor of electrical and computer engineering at the University of Toronto, led a team that built a prototype AI worm using freely available AI models.

Papernot, who has signed an open letter with recommendations for policymakers, spoke to AFP about why AI makes these attacks harder to stop.

The interview has been edited for length and clarity.

Q: How is AI changing these types of attacks?

A: The reality is that a lot of these "IoT" (Internet of Things or connected home) devices have very poor cybersecurity hygiene, and so they are very easy targets for bad actors to compromise.

Historically, a worm would exploit a fixed set of vulnerabilities, and that meant you could fix these vulnerabilities and prevent the worm from propagating any further to devices that had been updated.

Now, with AI, an attacker can design a worm that tailors the attack strategy to that device's specifics and these vulnerabilities can be different.

This makes it significantly harder to interrupt the propagation of an AI-driven computer worm, because you have to anticipate all the possible vulnerabilities the devices might present.

Some of these vulnerabilities are not software bugs.

They can also be misconfigurations -- for instance, if someone has used the same password for multiple services, or hasn't configured a database properly.

What AI-driven malware does is accelerate the pace at which these vulnerabilities are exploited and the scale.

In the past, your corner coffee shop, or the laptop in your basement that you use very rarely, might not be up to date, but it wouldn't be targeted by a human attacker, because they don't have the resources and the time to do that.

Now, the autonomy gained by using AI means that all these devices that are less immediate targets become interesting targets as a launchpad to attack more interesting targets.

But we have to keep in mind that a human is launching an agent.

Once launched, the agent can replicate, as we showed in our work, but it is launched by a human.

Q: What can we do to mitigate the risks?

A: We won't bring the risk to zero and we have to realize that, despite the fact that some of these vulnerability fixes are available, people don't update their devices or don't apply the fixes in a timely manner.

Despite cybersecurity experts calling for increased cybersecurity for decades, we still live in a world where a lot more can be invested in improving the security of our computer systems.

There are really three things we recommend in the very short term.

The first is to change development and testing practices to ensure that AI models are initially disconnected from the internet, as a precaution, to understand their behavior better before we expose critical infrastructure on the internet [to them].

The second recommendation is to create an international body to standardize both how we train and how we contain AI models. Both are important. We want to train models that are safer.

The last part is about making reporting of security incidents mandatory. Right now, we are only getting information as a function of what AI developers are willing to release, and that's a big problem.

In particular, the academic community needs a lot more information about incidents to be able to develop countermeasures much more quickly.

Otherwise, we have to do a lot of research on understanding the threat first, before we can develop the countermeasures. Otherwise, we won't see what bad actors are doing until an attack is launched.



Impossible Metals Says It Has Successfully Tested AI Seabed Mining Robot

Impossible Metals tested an autonomous AI robot that retrieved about 450 grams of polymetallic nodules from more than 4,000 meters deep in the Pacific Ocean in September. (AFP)
Impossible Metals tested an autonomous AI robot that retrieved about 450 grams of polymetallic nodules from more than 4,000 meters deep in the Pacific Ocean in September. (AFP)
TT

Impossible Metals Says It Has Successfully Tested AI Seabed Mining Robot

Impossible Metals tested an autonomous AI robot that retrieved about 450 grams of polymetallic nodules from more than 4,000 meters deep in the Pacific Ocean in September. (AFP)
Impossible Metals tested an autonomous AI robot that retrieved about 450 grams of polymetallic nodules from more than 4,000 meters deep in the Pacific Ocean in September. (AFP)

Deep-sea mining firm Impossible Metals said on Monday it has successfully tested an autonomous robot that uses AI to selectively retrieve critical minerals at depths of more than 4,000 meters in the ‌Pacific Ocean.

Parts of the ocean seabed contain potato-shaped rocks known as polymetallic nodules filled with copper, nickel and other minerals used in electric vehicles and electronics, reported Reuters.

Many companies in the nascent industry plan to use a vacuum-like device to retrieve those nodules, sparking controversy from environmentalists.

Impossible Metals, which is ⁠backed by Y Combinator, has developed a robot with a large claw that it says uses artificial intelligence to avoid aquatic life and reduce environmental impacts.

The test was part of a scientific expedition in September.

The robot directed itself to the seabed, retrieved about 450 grams of nodules and then returned to the surface.

Results from the test will be used to develop a larger version of the robot. The company still needs regulatory approval from the International Seabed Authority or the US government.

Separately, Impossible Metals said that the family of FedEx founder Fred Smith had participated in the company's Series A funding round, which closed earlier this year.

The company declined to provide the value of the round, but media reports put it near $25 million, ⁠an amount that values the company at roughly $110 million.

Richard Smith, a FedEx executive and vice chairman of FWS Enterprises, the Smith family office, said it was Impossible's technology that ⁠drew him in.

“It very responsibly, in an ecologically friendly way, goes down there and plucks these nodules off the seabed floor without disturbing the environment,” ⁠said Smith, who declined to provide the value of the investment but said it was “substantial.”

The Smith family is now Impossible's second-largest shareholder.


Nvidia Heads for $400 Billion Two-Day Wipeout as CEO Affair Fallout Deepens

The Nvidia logo outside Nvidia's Shanghai office building in Shanghai, China, November 10, 2026. (Reuters)
The Nvidia logo outside Nvidia's Shanghai office building in Shanghai, China, November 10, 2026. (Reuters)
TT

Nvidia Heads for $400 Billion Two-Day Wipeout as CEO Affair Fallout Deepens

The Nvidia logo outside Nvidia's Shanghai office building in Shanghai, China, November 10, 2026. (Reuters)
The Nvidia logo outside Nvidia's Shanghai office building in Shanghai, China, November 10, 2026. (Reuters)

Nvidia shares extended losses on Tuesday, putting the chip giant on track to shed about $400 billion in market value over two sessions as concerns mounted over the fallout from CEO Jensen Huang's affair with a Chinese businesswoman alleged to be a spy for Beijing.

The CEO's relationship with Alice Huang helped facilitate shipments of at least $7.5 billion worth of Nvidia chips to Chinese firms, The Information reported on Tuesday, citing two people who know her. Alice Huang and Jensen Huang are not related.

The allegations have intensified scrutiny of Nvidia's governance, export-control compliance and reliance on Huang's leadership, adding to investor unease over whether massive spending on artificial intelligence infrastructure ‌will deliver adequate ‌returns.

Nvidia shares fell more than 5% after losing 4.3% ‌on ⁠Monday, when the company ⁠shed $192.08 billion in market value. Tuesday’s decline put another roughly $221 billion at risk if losses hold.

"The investment debate is whether this remains company-specific or triggers restrictions and disruption across AI infrastructure," Bank of America analysts wrote in a note. "We would not equate allegations with proven business damage."

Nvidia's board launched an internal investigation on Sunday, while the FBI is examining whether the 63-year-old Huang was targeted by a ⁠foreign intelligence operation. Huang has told employees he was unfaithful ‌to his wife but provided no ‌details about the relationship.

Alice Huang, also known as Huang Le, along with her husband and sister, ‌built a network of companies across China and Singapore to acquire Nvidia ‌chips, The Information reported.

At the center of the network was Shanghai Century Litong, a China-based company that secured preferred cloud partner status with Nvidia in 2023, giving it priority access to the company's chips, the report said.

When orders for Nvidia chips subject to ‌US export restrictions stalled, Alice Huang used ties to Nvidia and its distributors to secure supplies, helping move about 240,000 H100, ⁠H200 and some ⁠A100 chips into China, according to the report.

Century Litong rented computing capacity and operated data centers using Nvidia chips, including some barred from sale to China. Some of the chips ultimately served Alibaba, Tencent and smaller Chinese AI labs, The Information reported.

Alice Huang also held an executive position at Singapore-registered Megaspeed, which was linked to Century Litong and has been investigated by US authorities over Nvidia chip shipments from Southeast Asia to China, the report said.

The Information reported that Alice Huang leveraged her relationship with Jensen Huang to bolster her credibility while keeping her role vague and avoiding a paper trail.

Nvidia did not immediately respond to a Reuters request for comment. Alibaba, Tencent, Megaspeed and Shanghai Century Litong were not immediately available for comment outside regular business hours. Reuters could not independently verify the report.


Quantum Could Solve AI Energy Drain, Says Tech Founder

Founder of Australian startup company Diraq, Andrew Dzurak, is reflected in a panel housing a cooling system for a quantum chip and computer, as he poses for a photograph at a data center located in Sydney on November 9, 2026. (AFP)
Founder of Australian startup company Diraq, Andrew Dzurak, is reflected in a panel housing a cooling system for a quantum chip and computer, as he poses for a photograph at a data center located in Sydney on November 9, 2026. (AFP)
TT

Quantum Could Solve AI Energy Drain, Says Tech Founder

Founder of Australian startup company Diraq, Andrew Dzurak, is reflected in a panel housing a cooling system for a quantum chip and computer, as he poses for a photograph at a data center located in Sydney on November 9, 2026. (AFP)
Founder of Australian startup company Diraq, Andrew Dzurak, is reflected in a panel housing a cooling system for a quantum chip and computer, as he poses for a photograph at a data center located in Sydney on November 9, 2026. (AFP)

Quantum computers could curb the vast energy demands of artificial intelligence data centers with their ability to run complex calculations, one of the tech's pioneers told AFP.

Australian startup Diraq is conducting the first field test worldwide putting a small quantum computer based on a silicon chip in a commercial data center in Sydney.

Gigawatt-scale energy demand from large data centers -- warehouse-like buildings full of servers used to train or run AI tools -- requires new power plants or increases pressure on electrical grids, sparking community backlash.

But, according to Diraq founder Andrew Dzurak, quantum computing is making the leap out of university labs and will be ready to lift the burden in data centers within five years.

"One of the reasons the next generations of AI are looking to consume so much power is that a whole class of the problems they are solving are problems that really are much better suited to quantum calculations," he told AFP.

Quantum computers harness the power of quantum mechanics -- laws that govern how physics operates at atomic and sub-atomic scales -- to process information in ways traditional computers cannot.

They usually require extremely cold cryogenic temperatures.

But Diraq's model, a self-contained system, draws less than 20 kilowatts of power as it runs alongside a computer humming at room temperature in the data center.

Potential uses beyond powering AI include rapidly and energy-efficiently calculating the "behavior of molecules that could be used in pharmaceuticals, or much lighter batteries or the chemical industry", Dzurak said.

Using quantum for drug trials could boost success by 10 percent and reduce the cost of bringing a new medicine to market by a third, if it runs for under $100 million per year, the company estimates.

"It will only happen if we can build quantum computers that are cost-effective," Dzurak said.

- Australia's Silicon Valley -

Diraq is among six global companies at the third stage of the US Defense Advanced Research Projects Agency (DARPA) quantum benchmarking initiative -- which will test if any quantum computer can solve real problems at a cost that delivers value by 2033.

Another Australian startup, PsiQuantum, is also on the DARPA list and backed by a Au$1 billion Australian government investment to build a different model of quantum computer in Queensland state. Construction started in June.

The country is experiencing a Silicon Valley moment in quantum, said Dzurak.

Australian physicist Michelle Simmons' startup Silicon Quantum Computing was named by Time magazine as having one of the best inventions of 2026, a quantum-enhanced AI chip.

Australia has been a global leader in quantum science for decades, partly because of investment in startups emerging from universities, said President of the Australian Academy of Technological Sciences and Engineering Cathy Foley.

"Australia has more than 50 quantum-related businesses operating here," she told AFP.

But "commercialization is certainly one of our biggest challenges", she said.

"Australia is very good at making discoveries, but turning them into successful businesses requires much more than great science."

The Australian government has made quantum a priority through its AUKUS defense technology partnership with Britain and the United States, Foley said.

AUKUS has reduced the security barriers for sharing quantum technology between the countries, which has been "essential" for Diraq to work across Australia and its offices in the larger US market, said Dzurak.

Diraq's computer installed in Equinix's Sydney data center has eight quantum bits, or qubits, on a silicon chip. This will be gradually upgraded to 150,000 qubits by 2029.

Qubits, the basic unit of information, can store more data and make more advanced calculations faster than classical computers, using quantum mechanics principles.

Diraq plans to have a system with two million qubits on a silicon chip that reaches DARPA's definition by 2031, two years ahead of the deadline.

"We will have an energy usage of a couple of hundred kilowatts," he said.