CIOs eager to scale AI despite difficulty demonstrating ROI, survey finds

Security and AI

Among the IT leaders taking a cautious approach to AI is Saurabh Gugnani, the global head of cyber defense and application security at Dutch compliance firm TMF Group. “Adopting AI poses several security challenges, such as data privacy, attack vulnerability, and strict regulation compliance. Protecting sensitive data and ensuring the integrity of AI models against cyber threats, such as adversarial attacks, are key concerns for CIOs,” he said.

Additionally, traditional security measures often fall short of addressing the unique demands of AI technologies. To mitigate these risks, CIOs must implement AI-specific security protocols and conduct regular security audits, which take time, he added. “Crucially, framing comprehensive policies that govern AI use, ethical standards, and security practices is essential to safely integrate AI into organizational workflows, ensuring a secure and compliant adoption process.”

For Neeraj Kumar, CTO of Arkreach, the speed with which AI technologies are being developed itself poses a threat. “Even tech giants like Google and Meta sometimes rush to launch immature tech, leading to a wreckage of failed AI products. However, rushing to ship some underdeveloped AI solutions in your product pipeline because of FOMO (fear of missing out) is not a good idea. It can throw your entire delivery system into meltdown,” he said. “It is time to shift away from the launch first, think later, mentality and concentrate on creating the future by not repeating yesterday’s mistakes, even if it leads to a delay.”

Another challenge is the critical shortage of skilled professionals that, according to Gugnani, makes it challenging and costly for companies to hire and retain talent in machine learning, data science, and AI integration. “Upskilling existing staff to manage AI technologies requires significant time and financial investment. Many organizations hesitate to commit the necessary resources, slowing the integration of AI capabilities,” he said.

The ROI dilemma

IT leaders also face the ongoing challenge of demonstrating and calculating the return on investment (ROI) of technology initiatives. The Lenovo survey found that 61% of CIOs find it extremely challenging to prove the ROI of their tech investments, with 42% not expecting positive ROI from AI projects within the next year.

One of the main difficulties is calculating ROI to convince CFOs to approve budgets, and this challenge is also present when considering AI adoption, according to Abhishek Gupta, CIO of Dish TV. “Quantifying tangible benefits such as cost savings, productivity improvements, and top-line growth is relatively straightforward. However, calculating returns on softer aspects, such as improvements in user experience, can be challenging,” Gupta said. “Ultimately, AI initiatives should be viewed as a necessary business investment that can yield results over time as the portfolio of AI projects begins to deliver results.”