AI 2019

(Erhan Astam, Unsplash)

This article is brought to you thanks to the collaboration of The European Sting with the World Economic Forum.

Author: William Dixon, Head of Operations, Centre for Cybersecurity, World Economic Forum & Nicole Eagan, Chief Executive Officer, Darktrace


Cyberattacks are becoming ubiquitous and have been recognized as one of the most strategically significant risks facing the world today. In recent years, we have witnessed digital assaults against governments and the owners of critical infrastructure, large private corporations and smaller ones, educational institutions and non-profit organizations. Not only is no sector immune from cyberattacks, the level of sophistication of the threats they face is continually increasing.

The future of cybersecurity will be driven by a new class of subtle and stealthy attackers that has recently emerged. Their aim is not to steal data, but rather to manipulate or change it. There is little doubt that artificial intelligence (AI) will be used by attackers to drive the next major upgrade in cyber weaponry and will ultimately pioneer the malicious use of AI. AI’s fundamental ability to learn and adapt will usher in a new era in which highly-customised and human-mimicking attacks are scalable. ’Offensive AI’ – highly sophisticated and malicious attack code – will be able to mutate itself as it learns about its environment, and to expertly compromise systems with minimal chance of detection.

Prototype-AI attacks: a glimpse into the future

AI-powered cyberattacks are not a hypothetical future concept. All the required building blocks for the use of offensive AI already exist: highly sophisticated malware, financially motivated – and ruthless – criminals willing to use any means possible to increase their return on investment, and open-source AI research projects which make highly valuable information available in the public domain.

One of the most notorious pieces of contemporary malware – the Emotet trojan – is a prime example of a prototype-AI attack. Emotet’s main distribution mechanism is spam-phishing, usually via invoice scams that trick users into clicking on malicious email attachments. The Emotet authors have recently added another module to their trojan, which steals email data from infected victims. The intention behind this email exfiltration capability was previously unclear, but Emotet has recently been observed sending out contextualized phishing emails at scale. This means it can automatically insert itself into pre-existing email threads, advising the victim to click on a malicious attachment, which then appears in the final, malicious email. This insertion of the malware into pre-existing emails gives the phishing email more context, thereby making it appear more legitimate.

Yet the criminals behind the creation of Emotet could easily leverage AI to supercharge this attack. Currently, the message on the final phishing email is usually highly generic – “Please see attached”, for instance – and this may sometimes arouse suspicion. However, by leveraging an AI’s ability to learn and replicate natural language by analysing the context of the email thread, these phishing emails could become highly tailored to individuals. This would mean that an AI-powered Emotet trojan could create and insert entirely customized, more believable phishing emails. Crucially, it would be able to send these out at scale, which would allow criminals to increase the yield of their operations enormously.

The consequences of these developing attack methods could be highly destructive, and even life-threatening. By undermining data integrity, these stealthy attacks cause trust in organizations to falter, and may even cause systemic failures to occur. Imagine an oil rig using faulty geo-prospection data to drill for oil in the wrong place, or a physician making a diagnosis using compromised medical records. As the AI arms race continues, we can only expect this circle of innovation to escalate.

Offensive AI: a paradigm shift in cyberattacks

In 2017, the WannaCry ransomware attack hit organizations in over 150 countries around the world, marking the beginning of a new era in cyberattack sophistication. Its success lay in its ability to move laterally through an organization in a matter of seconds while paralysing hard drives, and the incident went on to inspire multiple copycat attacks. This cycle of “innovation” will continue, and attackers have already moved on to cryptocurrency mining malware, which secretly steals processing power to mine for digital currencies such as bitcoin, and banking trojans, a type of malware that steals financial data while masquerading as a genuine application.

The use of adversarial artificial intelligence will impact the security landscape in three key ways:

1 – Impersonation of trusted users

AI attacks will be highly tailored yet operate at scale. These malwares will be able to learn the nuances of an individual’s behaviour and language by analysing email and social media communications. They will be able to use this knowledge to replicate a user’s writing style, crafting messages that appear highly credible. Messages written by AI malware will therefore be almost impossible to distinguish from genuine communications. As the majority of attacks get into our systems through our inboxes, even the most cyber-aware computer user will be vulnerable.

2 – Blending into the background

Sophisticated threat actors can often maintain a long-term presence in their target environments for months at a time, without being detected. They move slowly and with caution, to evade traditional security controls and are often targeted to specific individuals and organizations. AI will also be able to learn the dominant communication channels and the best ports and protocols to use to move around a system, discretely blending in with routine activity. This ability to disguise itself amid the noise will mean that it is able to expertly spread within a digital environment, and stealthily compromise more devices than ever before. AI malware will also be able to analyse vast volumes of data at machine speed, rapidly identifying which data sets are valuable and which are not. This will save the (human) attacker a great deal of time and effort.

3 – Faster attacks with more effective consequences

Today’s most sophisticated attacks require skilled technicians to conduct research on their target and identify individuals of interest, understand their social network and observe over time how they interact with digital platforms. In tomorrow’s world, an offensive AI will be able to achieve the same level of sophistication in a fraction of the time, and at many times the scale.

Not only will AI-driven attacks be much more tailored and consequently more effective, their ability to understand context means they will be even harder to detect. Traditional security controls will be impotent against this new threat, as they can only spot predictable, pre-modelled activity. AI is constantly evolving and will become ever-more resistant to the categorization of threats that remains fundamental to the modus operandi of legacy security approaches.

Incorporating AI in the digital ecosystem

As we increasingly rely on connected systems and devices, we are quickly developing a highly advanced and heavily connected digital ecosystem. We will require partnerships and capabilities that prioritize winning the strategic battles that count – and safeguard not only economically valuable data held by the public and private sectors, but the confidence in digital systems that underpins social cohesion and democratic institutions.

Investment in new technology will play a critical role in this emerging reality and evolving ecosystem. According to Forrester’s Using AI for Evil report, “mainstream AI-powered hacking is just a matter of time”. Indeed, as we begin to see AI become part of the cyber attacker’s toolkit, the only way that we will be able to combat this malicious use of AI is with AI itself. Therefore, incorporating the technology into this ecosystem is crucial.

Counterattack: Fighting machine with machine

The cybersecurity community is already heavily investing in this new future, and is using AI solutions to rapidly detect and contain any emerging cyberthreats that have the potential to disrupt or compromise key data. Defensive AI is not merely a technological advantage in fighting cyberattacks, but a vital ally on this new battlefield. Rather than rely on security personnel to respond to incidents manually, organizations will instead use AI to fight back against a developing problem in the short term, while human teams will oversee the AI’s decision-making and perform remedial work that improves overall resilience in the long term.

AI-powered attacks will outpace human response teams and outwit current legacy-based defenses; therefore, the mutually-dependent partnership of human and AI will be the bedrock of defense strategies in the future. The battleground of the future is digital, and AI is the undisputed weapon of choice. There is no silver bullet to the generational challenge of cybersecurity, but one thing is clear: only AI can play AI at its own game. The technology is available, and the time to prepare is now.