image copyrightGetty Images
Service interruptions and computer failures are common, but not on this scale.
In the last week, Microsoft suffered a massive outage of its services; An apparent technical failure paralyzed all flights in the United States for hours, and the communication systems of the National Security Agency (NSA) had problems.
Not to mention the disruptions that affected major US banks, as well as a major flaw detected in a popular programming language.
These are isolated incidents, but the fact that they occurred so close to each other makes people wonder if they could be related.
There is no evidence to suggest this, but it does tell us a lot about how the modern technological world works and where the vulnerabilities lie.
What has been happening?
As is often the case, many service interruptions were the result of an update or installation failure somewhere in the system.
A Microsoft network update caused servers to become overloaded and shut down.
It was also improvements to an aging and outdated network that grounded thousands of planes in the United States over the weekend.
End of Most read
Problems with the NSA's internal communications system, and separately with Apple's iCloud and Apple Pay, were also attributed to glitches during system updates.
The interruption of payments at US banks is attributed to a "software glitch" at JP Morgan Chase. The bank claims that this is not a cybersecurity incident.
But some recent problems are, in fact, related to cybersecurity vulnerabilities.
Artificial intelligence was used to discover flaws in Java, the third most popular programming language in the world.
Researcher Scott Helme also discovered malicious code in Optimizely, a platform used by some of the largest companies in the world.
The hackers managed to steal around US$7,000 in cryptocurrency before being discovered and expelled.
Why are these things happening now?
All the news you need is on BBC World Service
Download it here
End of Ad app
Let's first address cyber incidents and the role of new AI models.
"I would be very surprised if all of these cyber attacks and the people perpetrating them were not using artificial intelligence to accelerate them," Helme told BBC Radio 4's Today programme.
Since the summer, AI companies like Anthropic and OpenAI have decided that it is too risky to give the general public access to the most advanced models before allowing certain companies and institutions to use them.
In these cases, security professionals can use the same technology to detect flaws themselves and alert others before criminals get to them.
This occurred in both the recent Java and PHP incidents, where there was a race against time to fix the vulnerabilities before they could be exploited.
And then there are the blackouts.
This is not new: the world stopped in recent years when CrowdStrike and then Amazon Web Services paralyzed devices and systems.
Part of this may be due to technical debt, a term used when operators opt for quick fixes for software rather than solving the underlying problem.
The accumulation of superficial solutions can end up causing more serious problems in the long term.
Since artificial intelligence is causing many companies to reevaluate their cybersecurity, they may release more updates to fix the vulnerabilities they have detected.
Service outages also often remind us of the importance of the physical technology that underpins our online world, such as cables, servers, and data centers.
In the air traffic control incident in the United States, for example, the problem arose when engineers were replacing outdated copper cables with fiber optics.
To what extent should we worry?
US Secretary of State Marco Rubio insists that the recent incidents are not related nor do they come from a single attacker.
"We are not under systemic attack," he said. "These events have nothing to do with each other, they are not interrelated."
In cases like PHP and Java, the "good guys" found the flaws, with the help of AI, before the "bad guys", allowing them to compile some fixes before the bugs could be widely exploited.
At the same time, Anthropic, among other leading AI companies, has warned about the risks posed by its own products.
Service outages, in particular, have an impact on the physical world, often causing disruptions to healthcare, schools, and people's ability to do their jobs.
Or, as happened over the weekend, the ability of planes to fly.
image source, Reuters
They also highlight the world's heavy dependence on a small number of actors: not only cutting-edge AI companies, but also infrastructure providers such as Microsoft and Amazon, which, along with Google, manage a large share of the world's computing servers.
In the PHP incident, the world held its breath as a programmer, working on a voluntary basis, struggled to release patches in a form that was easily applicable.
And Java, run by another tech giant, Oracle, is used by almost all of the world's largest companies.
"The Oracle Java vulnerability is a stark reminder of how heavily reliant businesses have become on a relatively small number of core technologies," said Sheila Pancholi, head of risk assurance at consultancy RSM UK.
"Organizations need to know where their critical technology dependencies reside, act quickly when vulnerabilities arise, and verify that their response has actually worked."
Meanwhile, Xander Van Raemdonck, a security engineer at Aikido Security, said the challenge is "knowing if you're exposed."
"Java comes included in a lot of enterprise software, so many companies don't have a complete list of where it runs," he noted.
Is this the new normal?
Maybe for now.
The AI tools used to help find security flaws are still so new that they will likely continue to find vulnerabilities that have been present in code for decades.
“AI is accelerating discovery on both sides and streamlining the amount of software being delivered,” said Martin Reynolds, field CTO at software delivery company Harness.
"Therefore, the number of programs that developers must repair and update is increasing, while the time they have to do the work is decreasing."
Technical debt will need to continue to be addressed, which means more system upgrades and greater potential for disruption.
Additionally, there is the constant threat of hackers, who are becoming more and more sophisticated.
Attackers only need to find one vulnerability to exploit, while defenders have to protect themselves against all possible cyber attacks.
"Attackers can move quickly and without worry, and experiment without risk of failure, so naturally they tend to move a little faster," Helme said.
"We are currently seeing a small gap between the movement speed of attackers and defenders."
The cat-and-mouse nature of cyberspace makes it unlikely to end anytime soon, and the impact of AI means the stakes are getting higher.
This article was originally written in English and we used an artificial intelligence tool to translate it. A BBC journalist reviewed the text before publication. Learn more about how we use AI.
Discover more
Video
End of Video




