AI: Threat or opportunity for cybersecurity?

In an era of increasingly sophisticated cybercrimes, artificial intelligence’s (AI) rapid involvement in the cybersecurity industry along with its associated opportunities and challenges has taken a mesial position on the table of discussions.

While AI has rendered network security, anti-malware, and fraud-detection software more formidable by identifying anomalies much faster than a human brain, it also poses a massive risk to cyber security. Cyber threats leveraging AI include brute force, denial of service (DoS), AI-generated deepfakes, and social engineering attacks.

 “As we embrace the use of newer technologies, cyber miscreants have also become smart enough to keep evolving. They monitor what is happening on the cyber security front and try to make use of these newer tools to evade detection or stay under the radar as much as possible,” shares Dr. Sanjay Katkar, Joint Managing Director, Quick Heal Technologies Limited.

For instance, nowadays, generative AI such as ChatGPT is used by cyber criminals to craft convincing phishing messages or launch a more complex attack that is easily evaded by traditional defenses that a nescient user may have.

 At the same time, cybersecurity firms have observed this pattern and are ensuring at their end to use advanced technologies to tackle the situation.

“We make use of AI which takes decisions in milliseconds to immediately block a malicious activity and then begins to do a deep dive on the cloud front in the back end. We are launching a new security solution soon which takes a step further and analyses the threat without affecting the user and alerts all deployed protection as well,” says Katkar.

With the number of daily threats going up significantly, automations like these have become necessary for data security. A recent report by cybersecurity firm Group-IB revealed that over 100,000 ChatGPT accounts have been compromised and their data is illegally traded on the dark web, with India alone accounting for 12,632 stolen credentials.

In fact, as AI tools become cheaper and more accessible, the risks to cyber security associated with the technology is expected to increase.

Data from cybersecurity firm CheckPoint Research shows that during the first quarter of 2023, India's average weekly attacks rose by 18% in comparison to the corresponding period in 2022, with each organisation facing an average of 2,108 weekly attacks per organisation.

Meanwhile, the number of bot attacks has also gone up by 48% (88 million attacks) in the second quarter as against the first one (59 million attacks), and nine out of 10 websites experienced a bot attack, as per the State of Application Security Q2 Report by Indusface, an application security software-as-a-service company.

Ultimately, the question is, in the realm of cybersecurity – is AI a threat or an opportunity?

Vishal Salvi, CEO at Quick Heal Technologies analogises the situation with the manufacturing of a gun – while it was deployed to enable protection and be used by defence personnel, it is also exploited by terrorists.

“Primarily, we encrypt the data so we can maintain confidentiality. But looking at the ransomware attacks that are happening, they are using the same tools of encryption to launch the attack. Our tools are being used against us,” says Salvi.

The cybersecurity industry, in defence, has developed various solutions to test the quality of a software which can be run to identify and exploit the vulnerabilities. However, attackers have figured out ways to misuse these technologies. For instance, Metasploit, a computer security project that provides information about security vulnerabilities, is one such tool which has become very popular amongst hackers.

However, not everything is as gloomy as it seems. Although the occurence of cyberattacks has increased, cybersecurity capability to block these attempts has also become more refined.

 While there has been a sharp increase of 90% in the frequency of cyberattacks against Indian websites during the second quarter of 2023 as compared to the first quarter, which witnessed 500 million attacks – on the other hand, the Indusface report also underlines that more than 947 million cyberattacks in India were thwarted in the second quarter of 2023, out of a total of 1.1 billion worldwide.

Salvi further asserts, “There are several detection technologies that just cannot be used by the hackers. These are only used by the good guys who protect and detect anomalies.”

Despite the several potential drawbacks, AI is expected to advance cybersecurity and assist organisations in developing stronger security measures.