The New World of AI and Cybersecurity

Shuman Ghosemajumder, MBA ’02, is most interested in technologies that impact the world on a large scale.

“Technology is what I view as the most leveraged way to be able to impact the world positively,” said Ghosemajumder at February’s MIT Sloan Alumni Online session. “And if you can create a new type of product that many people can use and benefit from, then it allows a small group of individuals to be able to create something that is used by the entire world.”

Ghosemajumder has applied this philosophy to his successful career in AI cybersecurity, developing products that protect individuals and companies from cyberattacks. He is the co-founder and CEO of Reken, an AI cybersecurity startup. Before that, he was the chief technology officer of Shape Security, which protected some of the world’s leading companies against sophisticated bots. Shape Security was acquired by F5, where Ghosemajumder then served as its global head of AI. He began his cybersecurity career at Google, where he was the global head of product for trust and safety.

While speaking with moderator Jackie Selby, EMBA ’21, Ghosemajumder shared his big-picture insights into cybersecurity and AI.

The evolution of cybersecurity

When Ghosemajumder began his cybersecurity career at Google, he and his team used large-scale machine learning models to protect Google’s advertising systems. At Shape Security, he found that cyberattacks were becoming more complex. Instead of hacking individual accounts, cybercriminals could now hack thousands of accounts at once through automation.

“One of the biggest things we learned a great deal about in that time period was how big of a threat automation can be,” Ghosemajumder remarked. “And that leads us to today, where we see that generative AI and many different AI techniques enable automation at a scale that we've never really imagined before.”

These automation technologies have required companies to change their approach to cybersecurity. Instead of focusing on controls that allow only certain users to access their virtual infrastructures and data, companies are now focused on monitoring account users’ behavior to identify any anomalies that could indicate a cyberattack.

Implementing these more complex cybersecurity strategies requires people with specific skill sets. This increasing demand for talent has led to a change in how cybersecurity is delivered. Instead of hiring people within their organizations to create cybersecurity, companies are now beginning to outsource this work to outside cybersecurity service organizations. AI technologies play an essential role in the delivery of cybersecurity, enabling companies and service organizations to execute cybersecurity measures with less staff.

AI and experimentation

Of course, AI technologies can also increase organizational efficiency outside of the cybersecurity space. The buzz around AI has led many companies to begin their own AI projects. Ghosemajumder advises companies to be intentional about which AI projects they take on. He has found that many companies have begun experimenting with AI without identifying a specific business problem that they are trying to use the technology to solve. This approach does little to benefit specific companies. At the same time, Ghosemajumder acknowledges that all AI projects can help us learn more about these technologies in general.

“In order to be able to figure out how you derive value from some of those new technologies, you have to have exploration and experimentation,” said Ghosemajumder.

Ghosemajumder is particularly excited about the experimentation that is happening around natural language and chatbots. Chatbots can quickly comb through large amounts of data to answer questions from customers and employees. This technology is especially useful in the health care industry, where it can be used to interact with patients and conduct research.

In addition to health care, Ghosemajumder touched on AI applications for the retail, entertainment, and marketing industries. He pointed to the retail industry as a good example of one that is applying intelligence augmentation to its work. Intelligence augmentation refers to the practice of using technologies to help human employees do their jobs more effectively rather than using them to replace human workers. In the retail industry, technologies can be used to analyze large data sets, and employees use this information to make decisions about organizational processes.

Looking ahead

When addressing the concern about whether AI will ultimately replace human workers and lead to unemployment, Ghosemajumder says that it depends on how quickly businesses integrate AI into job functions. If this integration happens incrementally, employees will have time to learn the new skills needed to perform job functions that have been changed by AI. This would help to prevent widespread job loss.

“What the next five to ten years look like will be very telling in terms of how quickly technology is going to change the nature of our jobs,” Ghosemajumder predicted.

Ghosemajumder is cautiously optimistic about the future of AI and cybersecurity. Humans will be interacting with robots and AI software multiple times a day through normal activities such as using their phones, driving their cars, and using electricity in their homes. As these technologies become more advanced, cybercriminals will also have access to more sophisticated technologies that they can use for large-scale cyberattacks. We will need to develop new high-powered technologies that can protect us against these new threats.

“I’m optimistic about creating mechanisms that allow us to be able to stay ahead of those risks,” said Ghosemajumder, “because ultimately this is what society wants.”

Register to attend next month’s MITSAO for a conversation with Ajit Kambil, SB '85, SM '89, PhD '93, Managing Director in Executive Accelerators at Deloitte LLP, on his recent book, The Leadership Accelerator.