Torras.ai

Today, we have the immense pleasure of speaking with Pilar, a true pathfinder in AI and technology. Pilar is the Senior Director of Engineering for AI Research Strategy at Google. She also serves on the boards of Eventbrite and Universal Diagnostics. Her career spans linguistics, entrepreneurship, and strategy, with major contributions to AI and conversational technologies. We’re excited to hear her insights and experiences. Welcome, Pilar!

Pilar’s Journey into AI

Jordi: You’ve built an impressive career spanning engineering, strategy, and leadership in AI. Could you share the key milestones of your journey?

Pilar: I never initially planned to work in AI. I started studying science and technology, with an initial focus on astrophysics. However, due to a sports accident, my academic path was interrupted, and I took a year off in the UK to learn English. That led me to study linguistics, specifically Anglo-Germanic philology.

While studying, I realized my passion for science and technology was unfulfilled. That’s when I discovered computational linguistics. I pursued a Master’s in Cognitive Science and Natural Language, fascinated by how the human brain processes language. That led to my PhD in Intelligent Dialogue Systems, which opened doors to entrepreneurship and AI research.

Eventually, I co-founded IndiSys, which focused on conversational AI. Later, I sold the company and joined Intel. Now, I work at Google, focusing on AI research strategy.

Entrepreneurial Competition and the Early Days

Jordi: Back in Spain, I was running a company called Inbenta, focusing on natural language processing. Competing against IndiSys was tough! I remember feeling frustrated losing projects to you.

Pilar: Yes, we were competitors, but there was also camaraderie. We were both startups in Spain, competing with much bigger companies in an emerging AI market.

Jordi: It’s amazing to see where we are now. I eventually sold Inbenta, and here we are discussing AI’s future!

The Impact of AI on Everyday Life

Jordi: How do you envision AI transforming everyday life in the near future?

Pilar: We’re already seeing the impact. AI is deeply embedded in our personal and professional lives. It influences markets, the news, and even how we feel about the future. AI is shaping the next phase of society.

Jordi: I remember my first experience with Google search. It was a transformative moment—like when I first tried ChatGPT. It’s a major paradigm shift.

Challenges in AI Safety and Governance

Jordi: What are the biggest challenges in making generative AI both accessible and safe for billions of users?

Pilar: One of the main areas of research right now is human value alignment. AI development must be responsible. These models are essentially black boxes—complex mathematical formulas predicting the next word or action. Ensuring fairness, safety, and transparency is critical.

The Evolution of Computational Linguistics

Jordi: Your background in computational linguistics is fascinating. How has it shaped your work?

Pilar: I’ve always bridged two worlds: science/technology and linguistics/humanities. Early AI models followed two approaches—engineering-driven and linguistics-driven. Now, AI relies heavily on data-driven methods, but we’re seeing a return to incorporating linguistic knowledge into models. Understanding language structure and cognition helps improve AI’s explainability.

Jordi: I imagine there’s huge demand for computational linguists now!

Pilar: Absolutely. AI development requires not just engineers and data scientists but also experts in linguistics, psychology, and humanities. AI is fundamentally about understanding and representing human knowledge.

Advice for Entrepreneurs in AI

Jordi: Many of our listeners are aspiring entrepreneurs in AI. What advice would you give them?

Pilar: Understand your strengths and weaknesses. AI is a tool—it doesn’t take away opportunities, it creates them. Educate yourself on AI fundamentals, as access to knowledge is greater than ever. AI enables people to achieve things faster and at a larger scale than before.

Also, think about where you fit in the AI value chain. You don’t necessarily need to build foundational AI models—there are opportunities in application layers, integrations, and specialized AI solutions.

The Future of AI

Jordi: What do you believe the future holds for AI?

Pilar: We’ve moved from needing programming languages to communicate with machines to using natural language. However, current AI still requires structured prompts. The next step is refining AI to understand intent without needing complex prompting.

The future is about AI agents—personal assistants that negotiate on our behalf, understand our habits, and optimize decision-making. AI will be more personalized, embedded, and intuitive.

The Importance of Asking “Why”

Jordi: One thing AI doesn’t do well is ask, “Why?” It generates outputs but doesn’t interrogate user intent.

Pilar: That’s an interesting point. AI should engage in more interactive questioning to refine user needs. It reminds me of the conversational agents we built years ago, where personality traits influenced their behavior. AI could evolve into a more proactive, personalized assistant.

Breakthroughs in AI and Science

Pilar: One of the most exciting AI applications is scientific discovery. AI-powered research is accelerating breakthroughs. Tools like AlphaFold are revolutionizing fields like medicine. I believe we’re at the start of another Renaissance.

Jordi: Maybe AI will finally uncover the elusive theory of everything!

How to Connect with Pilar

Jordi: For those who’d like to connect, what’s the best way?

Pilar: The best ways are LinkedIn and Twitter/X. Feel free to reach out!

Conclusion

Jordi: Pilar, thank you for this incredible conversation!

Pilar: Thank you! It’s been a pleasure.

Stay tuned for more exciting discussions on AI and its future!