Cybersecurity has always been a race between attack technology and defence technology. Today, the tech leading both sides of the race is artificial intelligence (AI). Companies are layering AI into their IT networks to secure their data in the cloud, while criminals are adopting ever-more sophisticated AI capabilities.
Criminal and non-criminal organisations use AI technologies such as machine learning, smart automation and virtual modelling in a similar way and for similar reasons. Both sides incorporate AI components into their existing apps and infrastructures to add visibility, insight and efficiency. Hackers, for example, can buy AI components via the dark web to modernise malware such as TrickBot, a six-year-old Trojan that now boasts smart automated capabilities and is many times more dangerous than it used to be.
But before we get carried away with nightmare tales of weaponised AI and automated data theft, let’s explore why, and how, emerging technologies such as AI are essential for protecting your data.
Most Popular Insights
Do you need help with AI Security?
Speak to one of our team about how we can help you today.
How AI keeps your cloud data safe
Any organisation reading about these nefarious new AI-powered threats may wonder whether it’s wise to allow sales data, business plans, customers’ personally-identifiable information (PII) and other sensitive data anywhere near the cloud. Especially now that companies face greater liability than ever for PII under GDPR and the Data Protection Act of 2018. There’s just too much at stake, particularly in sectors such as banking, healthcare and government.
We understand those fears. But at esynergy we firmly believe that an organisation must embrace a move to the cloud in order to be competitive and to provide the kind of service that customers expect. And rather than making your data more vulnerable, the cloud actually keeps it more secure.
This is because cloud providers enable businesses of all shapes and sizes to embed the very latest security technologies to guard their data while also unlocking its value. Services such as Microsoft Azure, Amazon’s S3 and Google’s cloud storage allow you to embed machine learning components, smart automation tools and other AI tech seamlessly into business operations and then scale them as needed. Once in place, those capabilities update instantly and automatically, so they always provide the latest benefits and protections.
Azure’s secure research environment for regulated data is a great example of how AI keeps data safe automatically, while freeing users to work with that data. Originally created for higher education institutions, the architecture can be used in any industry that requires data to be isolated securely for research, such as finance and medicine. The dataflow process is seamless and secure. When the client organisation uploads their data, Azure’s architecture automatically encrypts it, removes PII, creates a copy in a secure environment, deletes the original, allows privileged virtual desktop access, and then adds AI capabilities such as training the data set and managing machine learning models. Approved researchers are able to focus on their work with the data, while the AI security components continuously monitor the workload and its environment to discover and mitigate risks before they can do any damage.
Unlike legacy security tools, AI doesn’t have to be told what threats to look out for. Instead, it uses machine learning to automatically detect anomalies, and then mitigate threats before they get anywhere near your data.
“Our products use machine learning algorithms, trained on millions of malware samples, to identify threats that we haven’t seen before,” says Adam Kujawa, Security Evangelist at Malwarebytes. “There are also AI tools that help with network monitoring and log analysis, and can inform IT staff of a problem as soon as possible. The AI might miss the first attack, but then share that knowledge with other AI and learn from it, creating new ways to detect the new attack and so on.”
Kujawa also credits AI with speeding up and automating the detection of phishing attacks, social engineering attempts and malware infections. “If all is working well, the user won’t encounter threats at all, and the battles will be at lightning speed, computer versus computer.”
Why AI attack needs AI defence
Cyber criminals love emerging technologies. AI and machine learning add visibility to their operations (for example by revealing security holes), guide decisions (such as how and when to attack), and automate data fraud at a scale and speed that even the most energetic old-school con artist would never manage. These capabilities are not exactly hard to come by. There’s every chance that your own phone is equipped with AI sophisticated enough to create a deepfake that’d fool a CEO’s mum.
Worse still, gangs have sabotaged companies’ own automation systems to generate data fraud at industrial scale, and even sabotaged machine learning data sets to make them generate inaccurate or dangerous decisions. This so-called ‘AI poisoning’ can render cybersecurity systems useless, with grim implications for global security.
The best protection against AI attack is AI defence. AI and machine learning are the only scaling factors that can supervise these systems effectively in real time. The AI security tools embedded in many cloud providers’ services are more than up to the job of guarding your data from advancing threats, especially with their machine learning algorithms learning constantly to spot possible threats across your entire database and network.
“Artificial intelligence can spot the breadcrumbs of sophisticated attacks,” says Max Heinemeyer, VP of cyber innovation at security firm Darktrace. “It can autonomously interrupt the in-progress threats it detects at every stage, whether that be with a digital fake email created by AI to stealthy lateral movement, all without business disruption.”
AI is not only the easiest and most effective way to secure your data in the cloud, it’s also an essential defence against emerging threats – including attacks that make use of AI to achieve devastating scale. Find out more about how we can design a scalable, modern strategy to protect your data in the cloud.