Microsoft AI CEO Mustafa Suleyman to developers and researchers: Stop 'doing' projects

The air in the tech world is thick with the buzz of artificial intelligence. Every day, new research papers are published, new open-source models are released, and countless developers are "doing" AI projects. But according to Mustafa Suleyman, the new CEO of Microsoft AI, this frantic activity is missing the point. His message to developers and researchers is a clarion call: Stop "doing" projects that are merely technical exercises and start "building" solutions that serve a clear human purpose.

Suleyman, a co-founder of the pioneering AI lab DeepMind, is no stranger to the intoxicating allure of a technically brilliant model. He has been at the forefront of AI research during its most explosive period of growth. Yet, his perspective has matured from pure research to pragmatic application. He sees a landscape cluttered with impressive demos and clever hacks that, while intellectually stimulating, fail to address a fundamental question: What human problem does this solve?

The distinction he draws is critical. A "project" is often insular. It might be about achieving a new state-of-the-art score on a benchmark, fine-tuning a model for a marginal performance gain, or creating a fun demo that garners temporary attention on social media. These endeavors are rooted in the "how" of AI—the mechanics, the architecture, the code. They exist, in a sense, for the sake of AI itself.

"Building," in Suleyman's view, flips this paradigm. It starts not with a technical possibility, but with a human need. It forces the developer to ask "why" and "for whom" from the very beginning. This shift in mindset transforms the entire development process.

What does this look like in practice?

Instead of building a generic chatbot, the "builder" would focus on creating a compassionate, reliable, and accessible mental health first-aid tool for people in rural areas with limited access to therapists. Instead of training a model to simply recognize images, the "builder" would deploy it to help a small-scale farmer in Kenya diagnose crop disease from a smartphone photo, potentially saving their harvest and livelihood.

This philosophy is about moving AI out of the sandbox and into the messy, complex, and impactful real world. It’s about creating systems that augment human capabilities, alleviate burdens, and solve tangible problems in fields like healthcare, education, climate science, and governance. The metric of success is no longer just accuracy or precision, but adoption, utility, and positive human outcome.

Suleyman’s message is not a dismissal of pure research. Foundational research remains the bedrock upon which all applications are built. Rather, it is a challenge to apply the same rigor and ambition to the "last mile" of AI—the point where technology touches human life. He is urging the community to harness its incredible talent and energy not just for technical virtuosity, but for virtuous ends.

For the individual developer or researcher, this means pausing before starting a new project and conducting a simple but profound "human-centric" litmus test:

• Who is the user? Be specific.

• What problem are they facing? Be empathetic.

• How will this solution make their life, work, or community tangibly better? Be honest.

The future of AI will not be written by those who simply built the most elegant models, but by those who used this transformative technology to build a better, more equitable, and more human-centered world. Mustafa Suleyman’s leadership at Microsoft AI signals that the industry's most powerful players are now aligning behind this mission. The challenge, and the opportunity, is now in the hands of the builders.