In March 2016, the world witnessed an unprecedented moment in the evolution artificial intelligence (AI), when AlphaGo, a DeepMind AI, played against Lee Sedol – one of the greatest players of modern Go. The match reached a crucial juncture with Move 37 in Game 2, where AlphaGo made an unconventional and creative move that stunned both the audience as well as Lee Sedol.
Since then, this moment has been recognized as pivotal in the evolution AI. It was not just a demonstration of AI’s ability to play Go, but also a revelation that machines can think outside the box. This moment fundamentally changed the perception of AI. It went from being a tool that followed predefined rules into an entity capable to innovation. Since that fateful match in 2007, AI has continued to drive profound change across industries, from fraud detection to content recommendations. ChatGPT introduced generative AI, which demonstrated the power of AI.
Discover the latest in AI innovations. Join Microsoft at the NVIDIA AI Conference.
Learn more and register
ChatGPT’s critical moment
OpenAI’s release of ChatGPT in November 2022 was another important milestone in the evolution AI. ChatGPT is a large language-model capable of generating text that looks like it was written by a human. This model demonstrated the ability of AI to understand natural language. This capability opened new possibilities for AI applications ranging from customer service to creating content. ChatGPT was received with a mixture of excitement and awe, as the world realized that AI could transform how we communicate and interact with technology in order to improve our lives.
Rise of agentic AI
Today’s rise of agentic AI – systems capable of advanced reasoning, task execution and decision making – is revolutionizing how organizations operate. Agentic AI systems have been designed to achieve complex goals in a predictable and autonomous manner. Multi-modality is a productivity enabler that can effectively integrate humans into the loop. These systems can make contextual decisions and dynamically adjust plans to changing conditions. They can also take goal-directed action with minimal human oversight.
Deploy AI agentic today
Microsoft, NVIDIA and other companies are leading the way in developing and deploying AI agentic systems. They provide the infrastructure and tools that enable advanced capabilities. Azure AI Foundry, Azure OpenAI Service, and other tools are used to build AI agents capable of autonomously deciding, acting, and perceiving. These services allow the development of AI systems capable of performing more complex multi-step processes.
AI Agents and Agentic AI Systems: Microsoft has developed a variety of AI agents that automate business processes and work alongside or on behalf individuals and organizations. These agents are accessible via Microsoft Copilot Studios GitHub or Azure AI. They are designed to autonomously sense, decide and act in response to new circumstances. Azure AI-powered mobile data recorder copilots at BMW allow engineers to converse with the interface in natural language. This allows them to convert conversations into technical insights.
Multiple-agent systems () : Microsoft’s research and development on multi-agent AI has led to the creation modular, collaborative agents capable of dynamically adapting to different tasks. These systems are designed for seamless collaboration, improving overall performance and efficiency. Magnetic-One is a high-performance agentic system that can solve open-ended problems across multiple domains. This represents a significant advance in agent technology.
Collaboration with NVIDIA: Microsoft and NVIDIA collaborated across the entire technology stack including Azure accelerated instances with NVIDIA GPUs. This allows users to create agentic AI applications using NVIDIA GPUs, NVIDIA NIM Models, and NeMo Microservices, across their selected Azure Services, such as Azure Machine Learning or Azure Virtual Machines. NVIDIA NeMo Microservices also offer capabilities that support the creation and enhancement of agentic AI apps.
Beyond physical AI
The next wave of AI development will be physical AI. This is powered by AI models which can understand our world, interact with it, and generate their actions using advanced sensory input. Physical AI will open up a new frontier in digitalization for heavy industry, bringing intelligence and autonomy to warehouses and factories around the world, and driving major advances in autonomous transportation. Microsoft Azure now offers the NVIDIA Omniverse platform, which allows developers to create advanced applications for physical AI, simulation and digital twins that accelerate industrial digitization.
AI is evolving and promises to bring more profound changes to the world. The journey from a single Go board move to the emergence and development of agentic and real AI highlights the incredible potential that AI has to innovate, transform industries and elevate our everyday lives.
Experience AI innovation at NVIDIA GTC.
Discover cutting edge AI solutions from Microsoft & NVIDIA that push the boundaries for innovation. Microsoft will be at the NVIDIA AI Conference in San Jose from March 17-21, 2025. You can also join Microsoft virtually. Visit Microsoft’s booth #514 in San Jose, California to meet Azure and NVIDIA AI specialists and explore the latest AI hardware and technology. Attend Microsoft’s sessions and learn about Azure’s comprehensive AI platforms. Register today to learn more.
Microsoft and NVIDIA produced this content. This content was not produced by the editorial staff of MIT Technology Review.
