In October 2023, the British Library suffered “one of the worst cyber incidents in British history,” as described by Ciaran Martin, ex-CEO of the National Cyber Security Centre (NCSC).

What lessons can other organisations learn from the ransomware attack?

Read more in my article on the Exponential-e blog.

Read More

The debate over professionalizing software engineers is decades old. (The basic idea is that, like lawyers and architects, there should be some professional licensing requirement for software engineers.) Here’s a law journal article recommending the same idea for AI engineers.

This Article proposes another way: professionalizing AI engineering. Require AI engineers to obtain licenses to build commercial AI products, push them to collaborate on scientifically-supported, domain-specific technical standards, and charge them with policing themselves. This Article’s proposal addresses AI harms at their inception, influencing the very engineering decisions that give rise to them in the first place. By wresting control over information and system design away from companies and handing it to AI engineers, professionalization engenders trustworthy AI by design. Beyond recommending the specific policy solution of professionalization, this Article seeks to shift the discourse on AI away from an emphasis on light-touch, ex post solutions that address already-created products to a greater focus on ex ante controls that precede AI development. We’ve used this playbook before in fields requiring a high level of expertise where a duty to the public welfare must trump business motivations. What if, like doctors, AI engineers also vowed to do no harm?

I have mixed feelings about the idea. I can see the appeal, but it never seemed feasible. I’m not sure it’s feasible today.

Read More

Law enforcers have arrested nine suspected members of a prolific cyber-fraud gang

Read More

The genius at the heart of AI—its ability to sift through mountains of data, actually spot a needle in a haystack, and act on threats before they blossom into full-scale emergencies—it’s undeniable.

However, here’s the rub—every part of that impressive arsenal? It’s also up for grabs by the other side, and can (and will) arm them to launch attacks of unprecedented sophistication and elusiveness, the likes of which we’ve thankfully never seen up to now.

How do we wield this impressive technology to fortify our defenses, while preventing it from falling into the wrong hands? Can such a thing even be accomplished? Join me below as we take a closer look at how AI’s rapid rise is changing the landscape of cybersecurity.

AI as a Defense Tool

AI is a reliable navigator for charting the digital deluge—it has the ability to handle vast quantities of information rapidly on a level that no human could ever hope to match. It doesn’t take a huge leap to come to the conclusion that those capabilities can very easily be leveraged for defense.

Automated Threat Detection

Think of AI as the ever-watchful eye, tirelessly scanning the horizon for signs of trouble in the vast sea of data. Its capability to detect threats with speed and precision beyond human ken is our first line of defense against the shadows that lurk in the network traffic, camouflaged in ordinary user behavior, or embedded within the seemingly benign activities of countless applications.

AI isn’t just about spotting trouble; it’s about understanding it. Through machine learning, it constructs models that learn from the DNA of malware, enabling it to recognize new variants that bear the hallmarks of known threats. This is akin to recognizing an enemy’s tactics, even if their strategy evolves.

All of what I’ve said also here applies to incident response—with AI’s ability to automatically meet threats head-on making a holistic cybersecurity posture both easier to achieve and less resource-intensive for organizations of all sizes.

Predictive Analytics

By understanding the patterns and techniques used in previous breaches, AI models can predict where and how cybercriminals might strike next.

This foresight enables organizations to reinforce their defenses before an attack occurs, transforming cybersecurity from a reactive discipline into a proactive strategy that helps prevent breaches rather than merely responding to them.

The sophistication of predictive analytics lies in its use of diverse data sources, including threat intelligence feeds, anomaly detection reports, and global cybersecurity trends. This comprehensive view allows AI systems to identify correlations and causations that might elude human analysts.

Phishing Detection and Email Filtering

AI has stepped up as a pivotal ally in the ongoing skirmish against phishing and other forms of social engineering attacks, which too often lay the groundwork for more invasive security breaches.

Through meticulous analysis of email content, context, and even the finer points of metadata, AI-driven mechanisms have become adept at weeding out phishing schemes, easily recognizing the warning signs of identity theft attacks that could have easily slipped past older, rule-based spam filters. This includes picking up on the nuanced indicators of foul play hidden in an email’s text, layout, or the seemingly benign details about its origin.

AI in the Arsenal of Cyber Adversaries

But lest we forget, the very capabilities that make AI a formidable defender in our arsenal also open doors for those with malicious intent to turn these advanced tools against us.

Sophisticated Phishing Attacks

Gone are the days when phishing attempts were easily spotted by their clumsy, one-size-fits-all approach. Today, armed with AI, cybercriminals craft messages of deception tailored with a personal touch, drawing on a vast reservoir of data pilfered from breaches, social networks, and other digital footprints left online.

What’s more, AI’s capacity to automate and scale these attacks brings a level of efficiency and sophistication to phishing operations that were once beyond the reach of many attackers. For example, the use of generative AI to swiftly set up convincing websites or to deploy a QR code generator as a lure—these are but a few tricks in the modern phisher’s toolkit, making the digital waters even more perilous for the unwary.

Automated Hacking Tools

These tools can comb through networks and systems with an efficiency and speed that was once unimaginable, pinpointing vulnerabilities with precision. They’re not just fast; they’re smart, learning to recognize patterns and security lapses, and then suggesting how they might be exploited.

Now, individuals without deep technical expertise can launch complex attacks, a development that has made the cyber battleground more unpredictable and dangerous.

Evasion of Detection Systems

Traditional security defenses often play a game of catch-up, relying on known signatures or patterns to flag malicious activity. AI-driven threats turn this approach on its head, analyzing and understanding the mechanisms of detection to actively avoid them. This chameleon-like ability to adapt in real-time renders these threats incredibly difficult to spot and neutralize.

Malware deployed by these advanced adversaries can alter its behavior based on the defenses it encounters, sneaking through the cracks of security systems designed to stop yesterday’s threats.

Surveillance and Espionage

AI has also empowered cyber adversaries with sophisticated surveillance and data analysis capabilities. By automating the sifting and interpretation of information from a myriad of sources, including social media, public databases, and even the Internet of Things through simple smart devices you probably already have in your home, attackers can uncover a wealth of sensitive data.

For instance, imagine cyber criminals using AI to spy on a CFO of a SaaS online payments solution provider and learn his or her habits. Within a short time, they could easily get their hands not only on client information but also a backdoor into the company’s API.

Navigating the AI-enhanced Cybersecurity Landscape

To harness AI’s full potential in cybersecurity without falling prey to its pitfalls, organizations and individuals must adopt strategic approaches that balance innovation with caution.

One key strategy is developing AI models that are specifically designed to detect and counteract AI-driven threats. Additionally, it’s critical to continuously update and train AI systems with diverse data sets to protect against evolving cyber threats while avoiding biases that could undermine their effectiveness.

It’s also important to put ethical considerations at the forefront of AI development and usage in cybersecurity. This involves creating AI technologies that respect privacy, ensure fairness, and are designed with accountability in mind.

The ethical concerns of AI are not restricted to cybercriminals. Imagine a corporation like your insurance provider using AI to scrape the dark web for data about if you’ve taken out a loan against your house, if you’ve gambled, and a million other things. They might not tell you, but they’ll certainly use this data to bleed you dry.

Finally, as AI continues to transform the cybersecurity landscape, ongoing education and awareness will become even more critical. This includes training cybersecurity professionals in AI technologies and strategies to equip them with the skills required to defend against AI-driven threats. It’s also important to raise awareness among the general public about the potential risks and safeguards associated with AI in cybersecurity.

Wrapping Up

As AI forges ahead, its double-edged nature becomes increasingly apparent in the cybersecurity domain. On one side, it offers a beacon of hope for more intelligent, autonomous threat management. On the flip side, it arms adversaries with tools of unprecedented sophistication, introducing fresh vulnerabilities into the digital ecosystem.

Adopting a balanced approach based on ethics, accountability, transparency, and widespread education is our best bet in maximizing AI’s defensive potential while neutralizing or at least ameliorating its threats, ensuring a secure digital future for all.

Read More

Mandiant observes what it claims is the first ever APT29 campaign aimed at political parties

Read More