AI comes alive: From bartenders, to surgical aides, to puppies, robots of tomorrow are on the way

The science fiction of humanoid robots is no longer. Imagine a future where robots can not only work with us in factories, but also assist in surgeries and take care of our loved ones. Tesla plans to deploy thousands The age of humanoid robotics is closer than you think.

As more companies demonstrate groundbreaking innovations, this vision is becoming more tangible. The 2025 Consumer Electronics Show showcased several examples on how robotics are improving in terms of functionality and human-centric designs. These included ADAM, the robot bartender from Richtech Roboticsmixes over 50 different types of drinks, interacts with clients, and Tombot Inc.’s puppy dogs that wag their paws and make sounds to comfort older adults suffering from dementia. Although there is a market for the robots displayed at the show and others, it’s still early for this type of robotics to be widely deployed.

Despite this, real technological advances are being made. This includes “humanoid robots” that use generative artificial intelligence to create more humanlike abilities — enabling the robots to sense, learn and act in complex situations. From Optimus, by Tesla, to Aria. Realbotix will be a major player in the development of humanoid robotics over the next decade.

Despite these promising advancements, some experts caution that achieving fully human-like capabilities is still a distant goal. Yann LeCun, one of the “Godfathers” of AI, cited shortcomings in current technology. LeCun may be right, but that doesn’t mean that we won’t soon see more humanoid robotics. Elon Musk said recently that Tesla will be producing humanoid robots. He expects to ship several thousand Optimusunits in 2025. In 2026, there will be between 50,000 and 100,000 of these. This is a dramatic change from the handful of machines that exist today, performing circumscribed tasks. Musk is known to make mistakes with his timelines, as he did when he predicted the launch of the Tesla Model 3. He said in 2016 he expected to achieve fully autonomous driving within two years.

It is clear, however, that humanoid robotics are making significant progress. Tesla is not the only company pursuing this goal. Agility Robotics and Boston Dynamics are also leaders in the field of humanoid robots.

The Business Insider published recently I had a chat Agility Robotics CEO Peggy Johnson who said that it would be “very common” for humanoid robotics to work alongside humans in a variety workplaces. Last month, Figure announced that the number of robots in the workplace would increase to ten. LinkedIn post : “We delivered F.02 robots to our client and they are currently hard at work.” Figure, with significant backing from major investors such as Microsoft and Nvidia will provide fierce competition in the humanoid robotic market.

Creating a worldview

LeCun had a point. However, more progress is needed before robots can match human capabilities. It is easier to move parts around a factory than it is to navigate complex, dynamic environments.

The present generation of robots faces three key challenges: processing the visual information quickly enough to respond in real-time, understanding subtle cues from human behavior, and adapting unexpected changes in their environments. The majority of humanoid robotics are dependent on cloud computing, and the network latency that results can make simple tasks such as picking up an item difficult.

World Labs is a startup founded by Fei Fei Li, the “AI Godmother”that is working to overcome robotics limitations. Li, in an interview with WIREDsaid: “The computer brain is behind the cameras. The physical world is seen by the computers through the cameras. Understanding the physical structure and dynamics of the physical environment is necessary to turn that vision into reasoning, generation, and eventual interaction. This technology is called spatial Intelligence.”

GenAI powers spatial intelligence by allowing robots to map their surroundings in real time, much like humans, and predict how objects may move or change. These advancements are essential for autonomous humanoid robotics that can navigate complex real-world scenarios and have the adaptability, decision-making skills, and intelligence needed to succeed.

While the spatial intelligence relies upon real-time data in order to build mental maps of an environment, another method is to help a humanoid robotic infer the realworld from a still image. As explained in a Generative World Explorer, a pre-published paper ( ), uses AI to create an entire virtual world based on a single picture, simulating how humans make inferences regarding their surroundings. This capability, which is still in the research stage, will help robots make split-second decision or navigate new environments using limited sensor data. This would allow robots to quickly adapt to new environments they have not experienced before.

The ChatGPT moment is coming for robotics

As World Labs, GenEx and Nvidia’s Cosmos tackle the challenge of equipping humanoid machines with real-world adaptability as well as interactive capabilities. Cosmos is a set of AI “world-foundation models” that helps robots understand physics, spatial relationships and other aspects. GR00T allows robots to observe humans in order to learn from them. Together, these technologies allow robots to understand what to do as well as how to do it in a natural way.

The robotics industry is pushing to equip humanoid machines with cognitive and physical adaptability. GR00T would allow humanoid robotics to assist in healthcare by observing medical professionals and mimicking them, while GenEx could help robots navigate disaster zones using limited visual input. As reported by Investor’s Business Daily.Nvidia CEO Jensen Huang stated: “The ChatGPT for robotics is on its way.”

Google DeepMind is another company that is working to create physical AI model. Timothy Brooks is a researcher at Google DeepMind. This month,posted on X about the company’s plans to make large-scale gen model that simulates the physical world. These new physical world models will help future humanoid robotics better predict, plan, and learn from their experiences.

Humanoid robots will be a reality by 2025. In the near future, they will be focused on specific tasks such as manufacturing, disaster response and logistics, where automation is of immediate value. As technology matures, broader applications such as retail or caregiving will be possible. The progress in AI and mechanical engineering are accelerating the development of humanoid robotics. Accenture, a consulting firm, recently noted the development of a full stack of robotics software and AI models that are designed to create machine autonomy within the human world. In their ” The company’s 2025 Technology Vision report states: “Over this decade, we will see robots casually interacting with humans, reasoning through unplanned tasks and taking independent actions in any type of environment.”

Wall Street Firm Morgan Stanley estimatedthe number of U.S. robots that are humanoid could reach 8 million by 2040, and 63 millions by 2050. The company stated that long-term demographic changes creating labor shortages could help drive their development and adoption.

Building trustworthy robots

Beyond purely technical challenges, societal objections may also need to be overcome. Public skepticism may prevent the adoption of humanoid robotics, even when they have clear benefits. Humanoid robots must be viewed as trustworthy and people must believe that they are beneficial to society. As Researchers are looking at ways to make robots more relatable. Engineers in Japan, for example, have created a mask made from human skin cells that can be attached to robots. According to a report published last summer, According to The New York Times,the study’s researcher stated: “Human like faces and expressions improve human-robot interaction, improving communication and empathy, making robots better in health care, service, and companionship roles.” In short, a human-like appearance improves trust.

Human-like robots must not only appear trustworthy, but also behave ethically and responsibly in order to gain human acceptance. In public places, for instance, humanoid robotics with cameras may inadvertently gather sensitive data such as conversations or face details, raising concerns over surveillance. To mitigate these risks, policies that ensure transparent data practices are essential.

The next decade

Humanoid robots are expected to focus on specific tasks in the near future, such as manufacturing and logistics, or disaster response, where automation is of immediate benefit. These specialized roles will highlight their strengths in structured environments, while broader applications like healthcare, caregiving, and retail operations, will emerge as the technology matures.

Humanoid robots will have a profound impact on human interactions and social norms as they become more visible. These machines will not only perform tasks, but also integrate into society, forcing humans to learn new ways of interacting with technology. Their adoption could ease labor problems in aging societies, improve efficiency in service industries, but also spark debates on job displacement, privacy, and human identity in a world increasingly automated. In order to prepare for these shifts, we will need not only technological progress but also thoughtful societal adaption.

By addressing the challenges and leveraging humanoid robotics’ efficiency and adaptability, we can ensure that these technologies serve as tools of progress. It is not just the job of tech leaders and policymakers to shape this future. Everyone can contribute. Public participation is essential to ensure that humanoid robotics improve society and meet real human needs. Gary Grossman, EVP Technology Practice at Edelmanis the global leader of the Edelman AI Center of Excellence.