Humanoid robotics in everyday life: how AI-supported robots are changing care, industry and service

Is innovation primarily the result of the interaction between different technologies?

Dr. Jürgen Baier: Exactly. Humanoid robotics is becoming interesting because our everyday work and life can be supported by really smart systems for the first time. Even seemingly simple routines, such as making a coffee in the morning or getting things out of a cupboard, are extremely complex. If you try to program such tasks in the traditional way, you very quickly reach your limits given the diversity of human environments.

AI opens up new possibilities here, particularly through so-called generalization. Systems can learn to move and act intelligently in different environments. For the first time ever, this gives us a realistic chance of reliably implementing such tasks.

Are humanoid robots fundamentally different from traditional industrial robots?

Dr. Jürgen Baier: Yes, the key difference lies in the lack of need for absolute precision. In our everyday lives, we place objects in similar places, but not with millimeter precision. Traditional robots were dependent on precisely this precision.
Humanoid robots, on the other hand, can deal with fuzziness. They are designed to cope with environments in which objects are not positioned exactly. This is a fundamental paradigm shift.

What does "humanoid" mean to you personally?

Dr. Jürgen Baier: Many people automatically associate "humanoid" with a human-like appearance. I prefer a different definition: humanoid means finding your way in a human-centered environment.
Our world is designed for the human body - for our hands, our body size and our movements. Door handles, drawers and tools are precisely tailored to this. Humanoid robots are systems that can move in these environments and work with tools designed for humans.

Do humanoid robots necessarily have to look like humans?

Dr. Jürgen Baier: That is still an open question and is the subject of intense debate. A well-known concept in this context is the so-called uncanny valley effect. It describes how robots should either be very human-like, including facial expressions and gestures, or be clearly different from humans. Anything in between is often perceived as unpleasant or even creepy.
Acceptance plays a decisive role here. Three fingers may be enough for certain tasks, technically speaking. But whether this is accepted by humans is another question. The situation is similar with size: small robots appear cute, larger ones can reach higher but may appear intimidating.
We will probably see a wide range of shapes, sizes and designs. Two main trends are currently emerging: walking robots with legs and mobile platforms with wheels. It is too early to say which will prevail in the long term.

What fields of application do you see for humanoid robotics?

Dr. Jürgen Baier: I see great potential in industry, care, education and service. Applications are particularly effective where physically strenuous or repetitive tasks can relieve people - for example in production or care.
Robots could, for example, take over transportation tasks or support care workers with heavy tasks. This leaves people more room for interpersonal aspects that cannot be automated.

Does naturalness in interaction and appearance play a special role in these areas in particular?

Dr. Jürgen Baier: Absolutely. We have learned a lot ourselves in the development of our robot. In early versions, we integrated a very human-like face, including a camera in the eyes. The feedback was clear: it looked unnatural and rather alienating.
Together with product designers, we derived key guiding principles from this. One of them is honesty: a robot should be recognizable as such. This creates safety. At the same time, communication must be meaningful - via speech, gestures or displays. Which combination will prevail remains to be seen.

What exactly is behind the robot you developed, Pib?

Dr. Jürgen Baier: Pib stands for Printable Intelligent Bot and is an open source project. Our aim is to democratize humanoid robotics. Anyone can participate.
A key feature of Pib is its 3D printability. The required print data is freely available. With a little patience - about a week's printing time - you can produce all the components yourself. Non-printable components can be added or purchased from us. In this way, we give many people access to humanoid robotics.

How is Pib structured?

Dr. Jürgen Baier: Pib has an average human size; we are currently mapping the upper body. It has arms with hands that each have five fingers and six degrees of freedom, including opposable thumbs.
The head has a display in the eye and nose area with a touch function. This can be used to start programs. A camera is integrated in the mouth area, which Pib uses to visually and spatially perceive its surroundings. A slight smile creates a friendly but not exaggerated impression.

How do you ensure that humanoid robotics are used responsibly?

Dr. Jürgen Baier: Transparency is key for us. As an open source project, both the design and the software are completely transparent. The data flows are particularly important here - for cameras and microphones, for example.

This openness creates trust and allays fears. Pib is a transparent robot in the truest sense of the word.

How is safety implemented in concrete terms?

Dr. Jürgen Baier: We work very closely with the community, especially with educational institutions. Safety and acceptance have always been our top priority there.
Pib works with a very low operating voltage of 7.5 volts. The joints are designed in such a way that no fingers can be trapped. Assembly is also designed to be safe. Many of these improvements are a direct result of user feedback.

What skills should companies that want to get involved with humanoid robotics have?

Dr. Jürgen Baier: The most important things are curiosity and motivation. Anyone who has ever worked with construction kits or furniture instructions has good prerequisites.
Our documentation is deliberately kept simple, inspired by well-known construction and furniture concepts. With a little patience, an interest in technology and enjoyment of challenges, you can build your own humanoid robot step by step.

What will collaboration between humans and robots look like in the future?

Dr. Jürgen Baier: That is still intensive research work. But one thing is clear: the approach must be human-centered. Communication via language, gestures and demonstration - for example through imitation learning - will play a major role.
Humans will remain the decision-makers. They show, explain and give instructions, while the systems learn to take over tasks with just a few attempts. We are not there yet, but the progress is clear.

How important is programming knowledge compared to soft skills?

Dr. Jürgen Baier: Many of our customers are medium-sized companies without large IT departments. There, the expertise lies with the specialists on site.
This domain knowledge must be easy to transfer into processes without complex automation projects. Humanoid robots must be able to integrate quickly and flexibly into existing processes.

Are humanoid robots also suitable for small quantities and a high number of variants?

Dr. Jürgen Baier: Yes, this is precisely where their strength lies. A wide range of variants is typical for SMEs, but makes traditional automation more difficult.
For large series, formal automation is often more efficient. Humanoid robots show their advantages where flexibility is required.

Where will humanoid robotics be in five years' time?

Dr. Jürgen Baier: I am convinced that we will encounter humanoid robots in everyday life in five years' time: in the home, in industry and in service.
Regulation remains a major challenge. Currently, humanoid robots are still treated like heavy industrial robots. New approaches are needed here. From a technical perspective, however, systems that can perform substantial tasks will be available in two to three years' time.

If you could give a humanoid robot one ability - what would it be?

Dr. Jürgen Baier: Quite selfishly: household tasks. Fetching things from the kitchen, finding keys, running small errands. Simply being able to say: "Where are my keys?" - and then get them, that would be great.

The interview was conducted by Christoph Raithel, Management Consultant, Bayern Innovativ GmbH, Nuremberg.