Humanoids: Professor of Robotics Bruno Siciliano says, ‘The EU is focusing on human-machine interaction’

Let’s forget about the humanoid breaking Usain Bolt’s 100-metre record. ‘I don’t think that should be our focus,’ says Bruno Siciliano, professor of automation and robotics at the University of Naples Federico II, director of the Prisma Lab and winner of the 2024 Pioneer Award from the IEEE Robotics and Automation Society – one of the sector’s leading accolades. “Let’s instead try to develop a European model of physical artificial intelligence, one that is geared towards interaction with humans. A model that should not, without exception, demonise even the use of Chinese hardware.”

Siciliano, who also coordinates the launch of the start-ups emerging from Federico II University in the field of robotics, calls for a dose of pragmatism by highlighting the risks of a potential ‘bubble’ and dismisses the prospect of a European scenario featuring ‘dark factories’ – factories where humanoids could, in theory, operate independently, without any workers, and where it wouldn’t even be necessary to switch on the lights. ‘At present, there is no adequate return on investment; one of the best-known examples is BMW’s change of heart regarding its use of Figure AI humanoids at its plant in South Carolina. The fully anthropomorphic version, complete with legs, is not always productive in an industrial plant, due to factors such as cost, stability and agility of movement in certain spaces, as well as broader regulatory issues. It can be an unnecessary complication compared to a robot with a fixed base and perhaps two arms, or compared to a robot that moves on wheels.”

This does not, however, mean giving up on the development of a European supply chain – a topic which Siciliano, together with international experts from the euRobin network, discussed in early September at the European Parliament in Brussels, advocating for an AI-based Robotics Act with a budget of 30 billion euros which, modelled on the Chips Act, would aim for a 33 per cent share of the global market by 2035. During the demonstrations held there, the robots dismantled electronic devices, delivered a parcel in an urban setting and cleared a table. ‘These are scenarios of coexistence,’ explains Siciliano, ‘and even today, the most effective scenario for physical AI is interaction between humans and humanoid robots, not replacement. Take, for example, solutions in the fields of transport or logistics to minimise work tasks involving greater physical effort – the same objective that, in some cases, can be achieved through the use of increasingly advanced exoskeletons. Not to mention all the possible applications in the fields of health technologies and personal care.”

It is along this line, according to the lecturer at Federico II University, that efforts should be made to develop the Italian supply chain. Today, domestic humanoid manufacturers can be counted on the fingers of one hand, and the government is considering measures to support their growth, touting the issue of sovereignty in response to Chinese dominance. ‘It is true that Europe today is caught between the Chinese model, which is heavily subsidised by the state, and the American model, with start-ups that have a vastly more developed venture capital market, but we must believe in our own model, following the European tradition in which robotics technology has wide-ranging applications in social contexts and not necessarily – or at least not exclusively – in factories.”

His views on the Chinese threat are also clear and, in a sense, go against the grain. ‘If the technology is truly open, I see no obstacles to using Chinese hardware on which to then develop Italian – or indeed European – technology. Let me give you an example. At Federico II, we are working on the development of a robot dog that can assist the visually impaired or accompany people with mobility issues during visits to archaeological sites. Well, we started with hardware from a Swiss company, but the real added value – created by us – is the smart lead which, using sensors, regulates and directs the robot’s movements. If the underlying technology is open-source and poses no security risks, why should we prevent the development of a similar application based on a Chinese robot, which might cost ten times less?”