In an exclusive interview with Telecom Review at MWC26, Guillem Martínez Roura, AI and Robotics Program Officer at the International Telecommunication Union (ITU), shared why 2026 marks a critical inflection point for embodied AI, how standards are shaping responsible scale, and why closing the global skills and connectivity gap is essential to ensure robotics serves humanity.
From over 480 AI standards to the expansion of AI for Good and the new Focus Group on Embodied AI, the ITU’s message is clear: deployment must be inclusive, interoperable, and aligned with human values from day one.
Robotics is emerging as a major theme at MWC 2026. What feels fundamentally different this year compared to previous cycles of robotics hype? Are we finally at an inflection point for real-world deployment?
Currently, we’re moving from impressive prototypes to readiness for deployment. The question is no longer which robotics use cases are at hand, but rather what conditions are necessary for inclusive adoption at scale.
The ITU, together with UNU, published a report two months ago that was announced at Davos, and this report establishes and defines the five foundational elements for broad, inclusive, and safe adoption of AI and robotics technologies.
Those five elements are data quality, access, and governance; digital infrastructure and equitable access; AI skills and talent development; responsible AI policy; and digital ecosystem development.
Those five elements are necessary to achieve deployment at scale. And going back to your question, yes, we are reaching an inflection point, but not a universal one, not a homogeneous one, because we need to have these five elements in place.
That’s why the AI for Good initiative is the leading United Nations platform on artificial intelligence. It’s so important to bring all the community on board and ensure that the robots we deploy and develop align with solving global challenges.
The AI for Good initiative was created in 2017 by the International Telecommunication Union, the United Nations specialized agency for digital technologies, in partnership with 53 United Nations agencies and in partnership with the government of Switzerland.
We operate under three main pillars: solutions and knowledge, which is about all these autonomous robotic systems that we are talking about; the second one is skills and capacity, ensuring that we bridge the AI skills gap; and the third one is about standards and policy, ensuring that we move from all these different solutions and high-level principles to practical implementation through standards.
We’re seeing humanoid robots, autonomous systems, and AI-powered machines across sectors. Which industries are realistically closest to large-scale adoption and how is the ITU preparing for this development?
We are seeing many industries advancing very fast. This goes back to the five principles I mentioned earlier, because the industries moving the fastest are those that are rapidly closing that readiness gap. They are deploying standards, strengthening evaluation maturity, and doing so at scale. As a result, they are the ones deploying robots at the fastest pace. We see this clearly in the manufacturing industry, in logistics, and even in some healthcare settings.
The ITU has been playing a very important role in bringing the community together and ensuring that these developments align with human-centered use cases.
For example, robotics for elderly care, robotics in hazardous environments, or robotics for disaster response and management. To further advance this work, we recently launched a new international effort called the Focus Group on Embodied AI.
This focus group will meet for the first time online on the 6th and 7th of May this year. It was created to look closely at embodied AI technologies from different angles. First, to ensure harmonization of terminology when it comes to embodied AI technologies. Second, to develop a pre-standardization roadmap for these technologies. And third, to focus on testing and benchmarking for performance, usability, reliability, and safety of embodied AI systems.
In parallel, we are looking at high-impact use cases to assess technological readiness. Equally important is collaboration. This focus group will work closely with other study groups within the ITU, as well as with other standard-setting organizations, industry, and academia to ensure that the robots we deploy are interoperable, interconnected, and capable of operating effectively within heterogeneous robotic systems.
In the context of “embodied AI”, where intelligence operates in physical systems, how do large language models change robotics and what are the limitations of integrating LLMs into real-world machines?
We’re clearly seeing how we are embedding AI into robotic systems, and large language models are definitely influencing and giving better capabilities to robotics, particularly when it comes to the cognitive and interactive layer of the system.
But robotics is not only about the interactive element; we’re also looking at multi-modal perception. We’re looking into the decision-making of the robot system, execution through manipulation tasks, and locomotion. Then, of course, there’s the interactive and collaborative layer, and finally, there is the learning part.
All these different elements and capabilities combined are the ones that are currently being explored at the ITU through our standards work. We have developed over 480 standards on AI—some already published, others still under development—and we look into all these different aspects.
Through the Focus Group on Embodied AI, we’re going to look into these capabilities in detail, with the objective of ensuring readiness for deployment.
As robots become more capable and connected through 5G and beyond, how do we ensure their alignment with human values, and could unequal access to advanced robotics deepen the digital divide globally?
The real challenge is moving from high-level human values and principles to actually embedding those human values and principles into robotic systems from the start, by design, not as an afterthought.
This is where standards play a critical role. They help us move from high-level principles to practical implementation. Through standards, we enable testing and benchmarking of AI systems to ensure they truly reflect the human values we want to embed in them. And that’s exactly the work we’re currently focusing on.
More importantly, we are making sure that through the robotics and AI revolution, we do not deepen the digital divides that already exist. We still have 2.2 billion people unconnected around the world. The ITU’s priorities are universal, meaningful connectivity and sustainable digital transformation.
Under this AI skills and capacity pillar, and through the AI Skills Coalition, we aim to democratize access to AI and robotics skills everywhere in the world. A recent and very important example is the AI and robotics youth training program launched in New York in September 2025. This is a collaboration between the ITU’s AI for Good, the Giga initiative, Google, and the will.i.am Foundation to ensure that we train and democratize AI and robotics learning in five pilot countries.
We’re starting in South Africa, Ghana, Kenya, Côte d’Ivoire, and Nigeria. The objective is to first train the trainers—the teachers. Once they are equipped with AI and robotics curricula, they return to their schools and teach these components to their students. The target is to train 20,000 students by the end of 2026, leveraging the Robotics for Good Youth Challenge, the AI Skills Coalition, and the ITU Giga initiative.
