AI comes to life: From bartenders to surgical assistants to puppies, tomorrow’s robots will be born

Humanoid robots are no longer the stuff of science fiction. Imagine a world where robots not only work with us in factories, but also greet us in stores, assist us in surgeries, and care for our loved ones. Tesla plans to introduce thousands of cars Optimus robot by 2026the era of humanoid robots is closer than we think.

This vision is becoming increasingly concrete as more companies showcase breakthrough innovations. The 2025 Consumer Electronics Show (CES) showcased several examples of how robotics is advancing in both functionality and human-centered design. These include: adam robot bartender Rich Tech Robotics mixes over 50 types of drinks and interacts with customers. Tombot Inc. puppies It wags its tail and makes sounds to comfort elderly people with dementia. While there may be a market for these and other robots on display at trade shows, it is still early days for widespread deployment of this type of robotic technology.

Nevertheless, real technological progress is being made in this area. This increasingly includes “humanoid” robots that use generative AI to create more human-like capabilities, allowing robots to learn, sense, and act in complex environments. From Tesla’s Optimus to Aria realboticsthe next decade will see a proliferation of humanoid robots.

Despite these promising advances, some experts warn that achieving full human-like capabilities is still a long way off. Yann LeCun, one of the “godfathers of AI,” points out the shortcomings of current technology: I had an argument recently. AI systems “do not have the ability to plan, reason, or understand the physical world.” He added that we can’t build robots that are smart enough today because “we can’t make robots smart enough.”

LeCun may be right, but that doesn’t mean we won’t see humanoid robots soon. Elon Musk recently said that Tesla will produce it. thousands of optimus units says it will ship in 2025 50,000～100,000 This is a dramatic increase from the small number of individuals that currently exist and perform limited functions. Of course, Musk is known for getting his timelines wrong. he said in 2016 Fully autonomous driving will be achieved within two years.

Nevertheless, it seems clear that humanoid robots are making significant progress. Tesla is not alone in pursuing this goal; other companies such as Agility Robotics, Boston Dynamics, and Figure AI are also among the leaders in the field of humanoid robots.

business insider recently had a conversation Peggy Johnson, CEO of Agility Robotics, said it will soon become “normal” for humanoid robots to co-work with humans in various workplaces. The numbers published in a magazine last month are LinkedIn Post: “We have delivered the F.02 humanoid robot to a commercial client. We are currently working hard on it.” With significant support from major investors such as Microsoft and Nvidia, Figure is leading the way in the humanoid robot market. There will be fierce competition.

Creating a worldview

But LeCun has a point, as further progress is needed before robots can achieve more fully human capabilities. It is easier to move parts within a factory than through a dynamic and complex environment.

The current generation of robots faces three important challenges. One is to process visual information quickly enough to react in real time. Understand subtle cues in human behavior. And it can also adapt to unexpected changes in the environment. Most current humanoid robots rely on cloud computing, and the resulting network latency can make simple tasks like picking up objects difficult.

One company working to overcome the limitations of current robotics is World Labs, a startup founded by “AI Godmother” Fei Fei Li. talk to wiredMr Lee said: “A computer’s physical world is seen through a camera, and behind the camera resides the computer’s brain. To turn that vision into reasoning, generation, and ultimately interaction, we need to understand the physical structures, You need to understand the physical dynamics. And that technology is called spatial intelligence.”

Gen AI enhances spatial intelligence by helping robots map their surroundings in real time and predicting how objects will move or change, just like humans. Such advances are essential to creating autonomous humanoid robots that can navigate complex real-world scenarios with the adaptability and decision-making skills necessary for success.

While spatial intelligence relies on real-time data to build mental maps of the environment, another approach is to allow humanoid robots to infer the real world from a single still image. As explained in Published papersGenerative World Explorer (GenEx) uses AI to create detailed virtual worlds from a single image, mimicking the way humans reason about their environments. Although still in the research phase, this feature could help robots make instant decisions and navigate new environments with limited sensor data. This allows you to quickly understand and adapt to spaces you have never experienced before.

ChatGPT’s moment in robotics is coming

While World Labs and GenEx are pushing the boundaries of AI inference, Nvidia’s Cosmos and GR00T are tackling the challenge of equipping humanoid robots with real-world adaptability and interaction capabilities. Cosmos is a family of AI “World Fundamental Models” that help robots understand physics and spatial relationships, while GR00T (Generalist Robot 00 Technology) helps robots learn from humans as apprentices learn from masters. Allows you to watch and learn. Combining these technologies will help robots understand both what to do and how to do it naturally.

These innovations reflect a broader push in the robotics industry to make humanoid robots both cognitively and physically adaptable. GR00T has the potential to enable humanoid robots to assist in healthcare by observing and imitating medical professionals, while GenEx has the potential to help robots navigate disaster zones by inferring the environment from limited visual input. may allow you to move. According to reports Investors Business DailyNvidia CEO Jensen Huang said, “The ChatGPT moment for robotics is coming.”

Another company working on creating physical AI models is Google DeepMind. Timothy Brooks, a research scientist there, Posted This month’s X describes the company’s plans to create large-scale generative models that simulate the physical world.

These new physical world models will better predict, plan, and learn from experience all the fundamental functions of future humanoid robots.

robot is coming

In early 2025, most humanoid robots are prototypes. In the short term, the focus will be on specific tasks such as manufacturing, logistics, and disaster response where automation can bring immediate value. As the technology matures, broader applications such as nursing care and retail interactions will emerge in the future. However, advances in AI and mechanical engineering are accelerating the development of these humanoid robots.

Consulting firm Accenture recently highlighted the development of a full stack of robotics hardware, software, and AI models aimed at creating machine autonomy in the human world. their “”2025 technology visionIn the report, the company says, “Over the next decade, we will see robots casually and commonly interacting with humans, logically handling unplanned tasks, and taking actions independently in all types of environments.” “It will be.”

wall street companies According to Morgan Stanley estimates, The number of humanoid robots in the U.S. could reach 8 million by 2040 and 63 million by 2050, the company said, adding that in addition to technological advances, long-term demographic trends creating labor shortages changes could accelerate development and adoption.

Building reliable robots

Beyond purely technical obstacles, potential social opposition must be overcome. If these concerns are not addressed, public skepticism could hinder the adoption of humanoid robots, even in areas where there are clear benefits. To be successful, deployed humanoid robots need to be seen as trustworthy, and people need to believe that they can help society. as noticed by MIT Technology Review“If such a robot were to walk into your living room today, few people would feel warm and comfortable.”

To meet challenges with trust, researchers are looking for ways to make robots appear more approachable. For example, Japanese engineers created a face mask from human skin cells and attached it to a robot. According to a study published last summer, reported by new york timesthe study’s lead researcher said, “Human-like faces and expressions improve communication and empathy in human-robot interactions, making robots more effective in their medical, service, and companion roles. ”. In other words, a human-like appearance increases trustworthiness.

In addition to appearing trustworthy, human-like robots must also consistently behave ethically and responsibly to ensure acceptance by humans. For example, in public spaces, humanoid robots equipped with cameras can inadvertently collect sensitive data such as speech and facial details, raising concerns about surveillance. Policies that ensure transparent data practices are critical to mitigating these risks.

next 10 years

In the short term, humanoid robots will focus on specific tasks where automation brings immediate value, such as manufacturing, logistics, and disaster response. These specialized roles highlight current strengths in structured environments, but as the technology matures, broader applications such as healthcare, nursing care, and retail operations will emerge.

As humanoid robots become more visible in everyday life, their presence will profoundly impact and potentially reshape human interactions and social norms. These machines will not only perform tasks, but will be integrated into the fabric of society, requiring humans to navigate new relationships with technology. These deployments have the potential to alleviate labor shortages in an aging society and improve efficiency in the service sector, but may also raise debates about job losses, privacy and human identity in an increasingly automated world. There is also. Preparing for these changes will require not only technological advances but also thoughtful societal adaptations.

Addressing the challenges and leveraging the efficiency and adaptability of humanoid robots can ensure that these technologies serve as tools of progress. Shaping this future is not just the responsibility of policymakers and technology leaders, but a conversation for everyone. Public participation is essential to ensure that humanoid robots improve society and respond to real human needs.

Gary Grossman is Vice President of Technology Practice. edelman He is also the global leader of the Edelman AI Center of Excellence.