Babak Hodjat, Chief AI Officer at Cognizant – Interview Series

Babak Hodjat, Chief AI Officer leads AI Research Labs, a team of developers and researchers that is advancing the state of the art in AI, building differentiated AI features in Cognizant’s offerings and leading the company’s AI for good initiatives.

Babak is the former cofounder and CEO of Sentient, responsible for the core technology behind the world’s largest distributed AI system. Babak was also the founder of the world’s first AI-driven hedge-fund, Sentient Investment Management.

Babak is a serial entrepreneur, having started a number of Silicon Valley companies as the main inventor and technologist. Prior to cofounding Sentient, Babak was the senior director of engineering at Sybase iAnywhere, where he led mobile solutions engineering. Prior to Sybase, Babak was the cofounder, CTO and board member of Dejima Inc. Babak is the primary inventor of Dejima’s patented, agent-oriented technology applied to intelligent interfaces for mobile and enterprise computing—the technology behind Apple’s Siri.

Babak has published more than 50 papers in the fields of artificial life, agent-oriented software engineering and distributed artificial intelligence, and has 39 issued US patents to his name. He is an expert in numerous fields of AI, including natural language processing, machine learning, evolutionary algorithms and distributed AI.

Cognizant is a global professional services and IT consulting company that helps organizations modernize their digital infrastructure, implement emerging technologies such as AI, cloud, data, and automation, and realign business processes to drive agility and growth.

You’ve founded multiple AI companies, authored two books, and helped pioneer technology that influenced Siri. Looking back, what personal experiences or turning points most shaped your belief in AI as a tool for real-world impact?

My fascination with AI began early on in my academic years and has only solidified since then. A few key turning points included creating some of the first Agent-based systems at Dejima for programming consumer goods like your VCR, the work we did at Sentient Technologies, applying AI to complex problems like financial trading. The natural language technology I developed for Siri, which was also agent-based, was another turning point.

These real-world applications demonstrated to me that AI could move beyond theoretical constructs and deliver tangible business value. We’re now headed into an explosive period for the commercialization of AI technologies, particularly with multi-agent systems that will completely transform how enterprises operate, enabling the automation of many complex tasks.

Let’s talk about your latest project—this AI-powered system for land-use planning developed with the University of Texas. What inspired the development of this tool, and how does it represent a shift from theoretical AI to real-world policy impact?

The AI system for land-use planning we developed in collaboration with The University of Texas at Austin was inspired by the need to address complex environmental and economic trade-offs in global land use. The machine-learning driven framework uses the technology behind Cognizant Neuro AI Decisioning and is built on the Project Resilience platform.

Our joint research team aimed to help realize the United Nations’ sustainable development goals by creating a tool that could optimize land use for maximizing carbon storage, minimizing economic disruptions and preserving food supply and habitats.  Traditional approaches often overlook nuanced trade-offs, such as location-specific effects of converting cropland or grassland into forests. We took a different path that leveraged evolutionary AI, a computational approach inspired by natural selection. It was designed to explore thousands of policy scenarios, iteratively improving and balancing competing objectives for more effective, context-sensitive land-use strategies.

The system marks a significant shift from theoretical AI concepts to real-world policy impact. It integrates historical land use data going back centuries and carbon data with sophisticated evolutionary algorithms to provide actionable, optimized recommendations rather than abstract predictions.

Separately, we have also created an interactive tool that generates and evaluates climate policy scenarios using the En-ROADS simulator, helping decision-makers compare and customize action plans. The system lets legislators and decision makers simulate policy incentives and directly understand trade-offs, helping them select targeted, efficient interventions. The system’s integration with platforms like Climate Interactive’s En-ROADS simulator enables scaling of AI-driven climate policy optimization to broader audiences, highlighting AI as a practical partner in tackling real-world sustainability challenges.

Can you walk us through how the tool works from a user’s perspective? What kind of decisions can it support, and how does it deliver tailored recommendations for different regions?

From a user’s standpoint, the AI-powered tool functions as an interactive decision-support platform in a dynamic, data-driven environment that helps you make smarter decisions that work for climate goals.

Policymakers, legislators, and other stakeholders can explore various land-use strategies and their environmental and economic impacts. Users can simulate incentives – tax credits for landowners, for example – and observe how they might influence land-use changes to reduce carbon.

It supports a range of decisions tailored to different regions. For example, it can help you determine where to make land changes for the best results, how much land to convert (like turning farmland into forest), and what the pros and cons are of different land policies. To provide custom advice, it looks at global land use history and carbon data to suggest different approaches for different areas. It can account for region-specific characteristics, such as latitude and land type.

The En-ROADS simulator, powered by Neuro AI, can help policymakers and legislators experiment with various tradeoffs for reaching different climate goals.

Evolutionary AI has been described as the “secret sauce” behind this project. How does this approach work in practice, and why is it so effective for solving complex environmental and policy challenges?

The idea behind evolutionary AI is inspired by natural selection in biology. As a computational approach in the context of land-use planning, in practice, it mimics natural evolution to find clever solutions for complex environmental problems that traditional methods struggle with.

Instead of trying to program perfect land-use policies upfront, the evolutionary AI approach creates different policy models and tests each one in simulated environments with real climate and land data. It keeps the best-performing policies and “breeds” them together, adding mutations to discover unexpected solutions. It repeats this process over many generations, winnowing poor performers and keeping the best across hundreds or thousands of scenarios.

This works well for environmental challenges because it doesn’t get overwhelmed by multiple variables such as soil types, climate conditions and economic factors.

Land-use policy often involves competing goals—economic growth, carbon reduction, food security. How does your system handle these trade-offs, and what kinds of unexpected insights has it surfaced so far?

Our AI system was built specifically to handle competing goals like economic growth, carbon reduction and food security. It generates Pareto fronts (an engineering concept used in multi-objective optimization) that trade off carbon impact and land-use change for different locations.

The research team encountered several unexpected insights. For example, while conventional wisdom recognizes forests as good at storing carbon, the AI didn’t default to recommending maximizing forest coverage everywhere. Instead, it revealed important distinctions: replacing rangelands like deserts and grasslands with forest wasn’t as effective as replacing crop land with forest. Geographic location also proved crucial. Identical land-use conversions produced different results depending on latitude.

Prioritization was one of the AI’s most practical insights. Rather than spreading efforts evenly, it suggested concentrating major land-use transformations in strategic locations where they’d have the greatest impact.

Project Resilience aims to scale this kind of AI utility to address Sustainable Development Goals beyond climate—such as energy, health, and even pandemic response. What excites you most about the potential to expand this platform across domains?

What excites me the most about the potential of this platform is that we are showing that building AI in a way that is collaborative, accessible, and adaptable can lead to powerful solutions for addressing big global challenges. The Project Resilience platform is a fitting example of putting those three principles into action.  Decision-makers, data scientists, and the public can join in to develop AI tools and make more informed decisions for significant impact. We invite your readers to be contributors here.

The AI Lab at Cognizant is now a major driver of innovation, with dozens of patents and a billion-dollar investment strategy. How do initiatives like this one fit into your broader roadmap for applied AI at scale?

The land-use AI initiative aligns perfectly with our approach to applied AI at Cognizant, which focuses on solving high-impact, complex real-world problems rather than purely academic exercises. Evolutionary AI can handle the often-complex trade-offs found in business and policy decisions. Tackling climate challenges through an approach that balances economic, social, and environmental factors demonstrates how AI can deliver practical value while managing competing priorities.

The work also reflects our vision for developing AI that augments human decision-making rather than replacing it.

You’ve led AI efforts across startups and enterprises. What’s the key to ensuring technologies like evolutionary AI remain explainable and actionable—not just powerful—for governments and industry stakeholders?

One of the greatest strengths of evolutionary AI is that it doesn’t simply aim to determine optimal solutions but can reveal strategy alternatives that expand stakeholders’ understanding of what’s possible.

AI and data must serve decision-making, not just generate reports. Decision makers are drowning in analyses while facing increasingly complex choices. We need to shift our focus from simply providing insights and predictions to creating interactive decision-support systems that offer prescriptive solutions based on available data. This approach empowers you to navigate complexity and make better decisions that evolve as circumstances change.

Looking ahead, where do you see the biggest opportunities for evolutionary AI to drive impact outside of land use—whether in infectious disease control, renewable energy planning, or something else entirely?

Infectious diseases, renewable energy planning and food insecurity are all worthy areas where evolutionary AI can drive impact. A COVID-19 era initiative we worked on shows the potential. Through Project Resilience, we built systems that could simultaneously optimize for pandemic containment and economic stability, helping governments like Iceland make data-driven decisions about school openings.

With evolutionary AI, we’re finally addressing the most pressing global challenges in a fundamentally different way, one that can recommend concrete policies that balance competing priorities rather than producing one-size-fits-all solutions.

The power of evolutionary AI is that it can simulate thousands of policy combinations, keeping what works and discarding what doesn’t. And it isn’t just theoretical. We’re building interactive tools that put this capability in the hands of real decision-makers.

After decades in the AI field, you’ve seen hype cycles come and go. What gives you confidence that the current wave—especially tools like this—is finally delivering on AI’s long-standing promise to improve society?

True progress toward AI’s promise to improve society happens when we move beyond the hype cycle to build systems that enhance human decision-making in areas that matter, including sustainability. That’s the real test of whether AI is finally delivering on its promise.

What I’ve consistently observed is that technology advances through predictable patterns of optimization and democratization rather than singular over-hyped moments. Look at computing history. We went from room-sized computers to powerful watches through continuous refinement, not one dramatic leap.

I do believe we’re at an inflection point where the technology can genuinely improve society. The path forward is through practical applications that solve real human problems, like our land-use work balancing carbon reduction with other goals. That’s how AI fulfills its promise: through measurable impact on our most challenging problems.

Cognizant recently set a Guinness World Record with the world’s largest vibe coding event—engaging over 53,000 employees across 40 countries and producing more than 30,000 prototypes. From your perspective, what does this say about the role of vibe coding in democratizing AI fluency in large organizations?

The scale and impact of our vibe coding event speaks volumes about how transformative this approach can be in democratizing AI fluency. Over 40% of participants were non-coders, and 20% had never written a line of code before. This says that vibe coding isn’t about lowering the bar, but about opening the doors to an evolving workforce with AI. We even used a multi-agent AI system to judge the 30,000 prototype entries in just a day, which would have taken a human team a whole year.

Instead of requiring deep programming expertise, vibe coding allows anyone with an idea to express it in natural language and collaborate with AI to bring it to life. For experienced developers, this lets them automate more of the mundane coding process, freeing them to focus more on higher-value work that drives business value. Personally, I was amazed at how quickly I could translate a complex algorithm from pseudo-code into a working application, freeing me to focus entirely on the creative and strategic aspects.

By lowering the barriers and enabling hands-on experimentation with generative AI, we’re moving AI fluency from a specialized skillset to a shared organizational capability. For organizations, vibe coding helps accelerate creativity, removes barriers, and unlocks the collective intelligence of an entire workforce at an unprecedented scale.

Beyond the impressive scale, what were some of the most meaningful outcomes of this vibe coding initiative? Do you see this as a template for how enterprises can cultivate innovation and applied AI skills at a global level?

The real impact of our Vibe Coding initiative was in the enthusiasm we saw across our organization, with employees from HR, sales, engineering, finance, legal, marketing, and more embracing AI and participating. Their ideas, grounded in deep domain knowledge and practical business insight, led to thousands of prototypes that might otherwise never have surfaced.

With more than 30,000 unique projects resulting from this effort, we are setting the pace for the AI economy, where everyone has the tools to innovate with AI. At the same time, we’re unlocking creativity at scale and empowering the workforce – both inside Cognizant and for our clients across the industries we serve – to become more AI literate.

We absolutely see this initiative as a template other organizations can replicate. By combining accessible AI-driven tools, a collaborative and open culture, and scalable AI-enabled evaluation, we can better serve our clients to unlock untapped creativity and accelerate the development of AI skills across their workforce.

Thank you for the great interview, readers who wish to learn more should visit Cognizant.