AI For Cybersecurity: Maximizing Strengths And Limiting Vulnerabilities The rapid pace of technology development that for long was seen as a trigger for digital transformation has given birth to new and advanced security challenges

Every day, we hear and experience the exciting tale of how technology is changing human lives. The rapid evolution of technology and the resultant innovation has improved access and safety standards and large-scale driven investment towards developing advanced technologies—including cloud, machine learning (ML) and artificial intelligence (AI). Across India, the government and the private sector have committed to invest in large-scale digital transformation and adopt new-age technologies such as AI and ML, already witnessing tremendous growth. As per a recent report from NASSCOM titled 'Unlocking Value from Data and AI' released in August 2020, the growth of AI will contribute significantly to India's economic and social development goals. AI data and AI hold the potential to add $450-500 billion to India's GDP by 2025.

But, this rapid pace of technology development that for long was seen as a trigger for digital transformation has given birth to new and advanced security challenges that could potentially imperil the interests of individual users, businesses, and governments across the world. For instance, the bad actors are now coming with more evolved and complex attack mechanisms that are difficult to intercept.

AI has today become the next battleground for cybersecurity, with industry experts having recognized AI as the technology that will determine the cybersecurity industry's future—creating both new risks and solutions. With the growing influence of AI in the IT landscape, the cybersecurity frontier's progress is increasingly under the radar. Cybercriminals are building advanced attacks using AI techniques to automate attacks and bypass the existing security infrastructure. With the flourishing culture of sharing and open-source technologies, bad actors can access the latest technologies that bring them head-to-head or even provide them with an edge over the host's defense mechanism.

Regardless of the origin, segment, size, or reputation, institutions and organizations across India are increasingly reeling under the threat of possible cyberattacks. As per the data collated by the ministry of electronics and information technology (MeitY), Indian citizens, commercial and legal entities, faced almost 700,000 cyberattacks till August this year.

Enterprises, too, are increasingly recognizing the need for speed and agility and working towards automating cybersecurity processes with next-generation firewalls that provide for built-in machine learning capabilities. The technology improvements help to provide an easy way for machines to spot any unruly network behaviour or pattern that could eventually emerge as a threat. Moreover, the fundamental deep machine learning processes would also help determine the origin of cyber-attacks, based on past threat analysis.

Concerning cybersecurity, AI has a much deeper role to play. It saves time by automating processes like monitoring the network, analysing user behaviour, patterns, and identifying any system irregularities. Besides, advanced AI solutions provide security professionals with additional resources and proactive alerts to defend the networks and data from cyber attackers. Moreover, with AI-based solutions, the capabilities keep improving as it learns from past behaviour to quickly identify any recurring security incidents and highlight/fix the vulnerabilities.

One of the significant advantages that machine learning and AI has brought to the fore is the detection capabilities and protection—especially in cases where the advanced and complex nature of threats sometimes is difficult to track by human professionals in the routine environment. But AI still lacks strength and more often fails in decidedly non-human ways. Most recently, an AI camera at a football game in Scotland kept mistaking the referee's bald head for the ball, ruining the game for fans. The AI-powered camera was programmed to follow the ball throughout the match. While this may have happened on the playing field, there is no opportunity to put AI in control without proper mechanism and proven records.

There is no doubt about the contribution that AI has made towards improving the existing cybersecurity solutions' capabilities and effectiveness, helping minimize the routine security risks. The debate is still on around how the transition will shape the future of cybersecurity. From identifying the recurring incidents and effectively highlighting the loopholes, and addressing the common vulnerabilities, AI has also provided a critical boost against tackling possible insider threats and managing the growing web of devices.

Given the nature of the technology, there would always be a risk of manipulation, and malfunction due to a variety of reasons including possible interference with codes and data from external means. Therefore, it's essential to understand the multiple dimensions and implications of the various AI-based actions. Considerable effort is required to implement the AI solutions righteously including dynamic standard-setting at both domestic and international levels for the purpose of testing, validating, and certifying AI tools to avoid any possible manipulation by advanced machine learning techniques. It's important that we don't rush into incorporating technologies but rather learn, understand, and invest in learning from the past—so we can maximize the progress and minimize the vulnerabilities to act against any future attack.