Artificial Intelligence (AI), or machine learning, is changing the technological landscape. As computers “learn” and make smarter “decisions”, response times become faster and efficiency and productivity increase.
In the world of cybersecurity, AI is playing a more active role in efforts to discover cyber attacks. But, in the wrong hands, can this type of technology do more harm than good?
Here is a look at the role of artificial intelligence in cybercrime.
On this side of the fence
Newer AI algorithms are very proficient at examining data traffic, access and transfer and identifying “outliers” or any abnormalities in data patterns. If something unusual is detected, the AI programs can take a deeper dive into the data to determine if there is a security breach in the system. For example, let us suppose a credit card is used regularly from a particular IP address, and that too likely twice in a month. And, if there is a sudden change in behaviour like a credit card is being used by some other IP or the excessive number of times than usual, it will alert the user for the abnormal pattern.
Another way that AI is used in the prevention of cyberattacks is through a method called “supervised learning”. With this method, the algorithm is provided with a number of inputs and outputs and over the time “learns” to detect threats by making judgements based on the data it sees or expects to see. For example, supervised learning may be used to detect advanced malware that is masquerading as a harmless piece of code.
Also, an AI-enabled VPN can detect if someone suddenly accesses a file with sensitive information, instead of regular files.
On the other side of the fence
On the flip side, experts suggest that AI can also be cause for concern in the field of cybersecurity.
As the complexity of human knowledge and machine learning grows, so will the threats against enterprises. Artificial Intelligence could be leveraged for nefarious purposes, such as the creation of ever-evolving malware or algorithms designed to infiltrate computer networks.
Several studies are currently underway to better understand the potential threats of AI in relation to cybercrime, and the results are concerning.
Potential threats may include:
- Theft of sensitive data from government and private organizations
- Hacked and weaponized drones
- Hacked and weaponized autonomous vehicles
- The exploitation of networks to interrupt workflow or steal data
- Social engineering attacks
- Disruption in the political arena, such as “misinformation” campaigns
So, when combined with human efforts, artificial intelligence can be utilized to bridge the gap in an organization’s cybersecurity workload. However, on the other side, if cybercriminals use AI, they can create deadly exploits and payloads.