​​Persuasive robotics: a field for experimentation in surveillance and security​

KEYS FACTORS OF THE ANALYSIS

• Recent research has demonstrated how the persuasiveness of robots can exceed that of humans when it comes to influencing certain decisions or changes in behavior. 

• Experimental studies have shown that people subconsciously consider that robots have some degree of inherent authority, which means that designing them to adopt or simulate human authoritarian roles may or may not be useful depending on the context or situation in question. 

• The level of authority and influence that a machine can achieve entails great potential opportunities for innovation in the field of surveillance, security or the protection of the home and physical spaces. 

• Progress with AI and data analytics are opening up a window of opportunities for robots to personalize interactions based on the information available in digital environments. However, these practices entail an ethical component for which regulatory adaptation is pending. 

The complexity of artificial intelligence and machine-learning systems, applied to the field of robotics and automation, is increasing by leaps and bounds each and every day . As machines become more precise in their ability to interact with humans, the objectives for which they are designed are becoming more diverse. 

One of the issues that has aroused most interest in the field of science is the persuasive capacity of robots. To what extent could a machine have the authority to influence human conduct? This area of research has resulted in the creation of the concept of persuasive robotics; in turn, this is serving as the basis for other research being performed aimed at understanding the potential of robots to change human attitudes or conduct.

The field of application of persuasive robotics is diverse, encompassing industries including health and education. However, without a doubt, one of the areas in which this trend has most potential is in the field of security and surveillance systems. 

TO WHAT EXTENT CAN A ROBOT ADOPT A ROLE OF AUTHORITY?

A study published recently by Science Robotics (1) intends to answer this question; the purpose of this research was to explore how a robot can adopt roles of authority and what impact this can have when it comes to modifying human conduct. 

Based on the results of this research, a robot can assume many roles when interacting socially with humans. These roles include functions that require some degree of authority, such as that of a nurse, teacher or guard and, although there is still a lot of work to be done to develop these capabilities, the results suggest that the machine's level of influence could be similar to that of a human when it comes to carrying out activities such as protection and security in the home. 

The Science Robotics study focused on comparing the impact generated as part of an interaction between a human and a machine in terms of formal authority. As part of the experiment, they used the well-known robot named Pepper (designed by Softbank Robotics), who tried to persuade a sample of people to change their responses in relation to a series of tasks that required paying attention and memorizing certain things. 

The results demonstrated that when the robot played a companion role (i.e. as if it were just another person participating in the study), it achieved a higher level of persuasion than when simulating a role of human authority. 

In general, this study demonstrated the high level of persuasive influence that robots have, as people potentially perceived them as a reliable figure and whose behavior was interpreted as dominant by the participants in the study. 

THE IMPORTANCE OF APPEARANCE AND SOCIAL CONTEXT

However, the experiment also demonstrated that the interactions generated some aversion among participants toward the authoritarian role of the robot, expressing that their ability to decide independently was being threatened. Furthermore, when the robot adopted an overtly authoritarian attitude, the participants perceived this as "threatening and inhumane".

THE PERSUASIVE CAPACITY OF HUMANS IS INHERENT TO ROBOTS, ESPECIALLY IN CERTAIN CONTEXTS AND ACTIVITIES

Therefore, the social context in which human-machine interactions take place is a decisive factor when it comes to producing a given result in terms of the robot's persuasive capacity. Thus, in certain contexts, the persuasive capacity of robots will be inherent to their condition as an automaton and, therefore, designing machines so that in this type of situation they adopt an authoritarian approach may not only be ineffective but also counterproductive. 

Persuasion strategies have been subject to intense study and a variety of techniques and tools have been developed that are known to influence conduct through emotions, rewards or penalties, authority, etc. All these strategies can also be implemented in a robot to use them just as a human would. However, the research mentioned here aims to understand how robots come with certain built-in mechanisms as standard that give them an authoritarian or influential role from the perspective of subconscious human perception. 

Therefore, the physical appearance and attitude of a robot may have to be different depending on the context of the interactions for which it is designed. Therefore, there will be important differences if they are dedicated to education, health or surveillance. 

IS IT POSSIBLE TO REPLICATE THE EFFECT OF HUMAN CONTACT?

Other recent research, by PLOS ONE (2), focused on trying to understand how humans respond to gestures and touch from robots, concluding that robots are more influential when they smile and touch people. 

As part of the experiment, 48 students interacted with the humanoid robot dubbed "NAO" (also owned by Softbank Robotics) as part of a conversation in which the robot acted as a janitor. In these situations, the robot adopted certain gestures and touched the other person's hand or refrained from doing so. The results revealed that students whose hand had been touched by the robot were much more likely to comply with the requirements or requests made by the janitor (in this case enrolling on a university course) than those with whom he had not established physical contact. Furthermore, students whose hand had been touched reported a better response to the interaction with the robot than those whose hand had not been touched.

PHYSICAL AND VISUAL CONTACT INFLUENCES THE IMPACT OF INTERACTIONS WITH ROBOTS ON HUMANS

The test therefore demonstrated that as is the case as part of interactions between humans, physical contact and gesticulation are also positively perceived when it comes to generating comfort, trust and influence as part of human-machine interactions. This opens the door to a wide range of possibilities in the field of persuasion, applying a much more precise, complex and specific type of contact. However, there is still much research to be done in this field with a view to understanding the complexity of physical human-robot interactions in depth, as these can be very different from interactions between humans. 

Another similar study, performed by the Eindhoven University of Technology (Netherlands) in Singapore (3), involved a sample group of individuals who the robot tried to persuade using two strategies, based on looks and gestures, to tell a story about the negative impact of lying. The results demonstrated that, when the robot maintained eye contact, the message gained more depth, regardless of its gestures or body expressions. 

The experiment thus demonstrated how, contrary to belief, a persuasive technique can be more effective when used alone than when used in combination with other persuasive techniques. Furthermore, it was found that eye contact between a robot and a human can have the same impact as eye contact between two humans.

SAME EXPECTATIONS AND SOCIAL STEREOTYPES

Another interesting aspect when investigating the persuasive power of robots are the expectations they generate in people. A study undertaken by the universities of Stanford and Harvard in 2000 (6) showed that people employ the same insights to machines as they apply to humans. To this end, if, for example, a robot is large and is threatening in size, it generates the same sense of authority as a human with the same physical characteristics. 

Secondly, the same biases and social constructs that apply to humans also apply to them. For example, a masculine looking robot generates different expectations than a feminine looking robot, and these expectations are based on the same biases that apply between humans.

PEOPLE JUDGE ROBOTS EMPLOYING THE SAME STEREOTYPES AS THEY USE TO JUDGE HUMANS

The study, performed by a public university in Singapore, used a robot guard to demonstrate how masculine features elicited a greater sense of authority among students. Forty students (20 boys and 20 girls) participated in this experiment, with the results demonstrating how biases applied equally to both genders and that both boys and girls saw a robot with masculine traits as having more authority. 

Various studies have demonstrated how social stereotypes are repeated as part of interactions between humans and robots, in particular in the field of gender differentiation. Another investigation performed by the Institute of Electrical and Electronic Engineers (10), using a robot that solicited donations from visitors at the Boston Museum of Science, found that men were more likely to donate money when the robot had female traits. And generally speaking, across both genders, everybody tended to identify robots with traits typical of the opposite sex as being more reliable and persuasive. 

ARTIFICIAL INTELLIGENCE AND BIG DATA

As part of this trend toward persuasive robotics, certain technologies are going to play a core role, mainly a combination of big data and advanced artificial intelligence. 

As is already the case with the algorithms employed by large technology platforms like Spotify (in the music industry) or Amazon (in the retail and e-commerce industry), robots can progressively integrate their own algorithms capable of analyzing the individual data of each person, thus personalizing their interactions with them, achieving a much more effective and contextualized capacity of persuasion. 

One example applicable to the surveillance and security sector would be as follows. A guard robot could, at some point in time, analyze data (from digital sources) pertaining to a specific individual who wants to access a facility and, depending on the results, adopt an attitude or employ one surveillance level or protocol or another, etc. This analysis can also offer information about the personality or habits of the specific individual (for example, based on their use of social media) with the robot using this to try to convince them or change their behavior. 

Having said that, this is a field with many ethical conundrums that still has to be regulated. Without looking further afield, the European Union is currently working to publish a regulatory AI framework that clarifies the extent to which organizations can process certain personal data belonging to citizens via artificial intelligence to establish profiles or segmentations that may incur bias, discriminate based on stereotypes or infringe upon the right to privacy. 

To this end, experts believe that, in order for this trend to continue developing, the owners of robots with persuasive characteristics will need to offer maximum transparency in terms of who is behind the robot's operations and what criteria are used. Furthermore, these humanoid machines must be designed to comply with current data protection legislation at all times.

Benefits and opportunities of persuasive robotics for the security and surveillance sector

• Robots can be used as tools to influence people's behavior, persuading them not to undertake certain actions. 

• In the field of domestic surveillance, for example, a robot could be used to persuade the tenants of a property to reduce their energy consumption. It could also persuade older people or children not to perform certain high-risk activities or to change their habits in terms of health, lifestyle, etc. 

• Robots can be programmed based on certain human features to cause the desired effect in each given situation. For example, using eye contact to build trust or to intimidate, etc. 

• Applying data analytics and AI algorithms, robots could individualize and contextualize each interaction with a human, achieving much greater persuasive power than a person could achieve. 

• Human guards could use robot companions to help them as part of their daily tasks, as well as to increase the effectiveness and scope of their security work.

BIBLIOGRAPHY

1. SCIENCE ROBOTICS. Persuasive robots should avoid authority: The effects of formal and real authority on persuasion in human-robot interaction. September 22, 2021 [accessed 1/7/21]. Available at: https://www.science.org/doi/10.1126/scirobotics.abd5186 

2. PLOS ONE. The persuasive power of robot touch. Behavioral and evaluative consequences of non-functional touch from a robot. May 5, 2021 [accessed 1/7/21]. Available at: https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0249554 

3. SPRINGER LINK. Making Robots Persuasive: The Influence of Combining Persuasive Strategies (Gazing and Gestures) by a Storytelling Robot on Its Persuasive Power. 2021 [accessed 07/01/21]. Available at: https://link.springer.com/chapter/10.1007/978-3-642-25504-5_8 

4. UC3M. RESEARCH PORTAL. Malfaz Vazquez, María Ángeles. 2022 [accessed 07/01/21]. Available at: https://researchportal.uc3m.es/display/inv21231 

5. CITESEERX. Machines and Mindlessness: Social Responses to Computers [accessed 1/7/21]. Available at: http://citeseerx.ist.psu.edu/viewdoc/downloaddoi=10.1.1.87.2456&rep=rep1&type=pdf 

6. SPRINGER LINK. When Stereotypes Meet Robots: The Effect of Gender Stereotypes on People's Acceptance of a Security Robot [accessed 1/7/21]. Available at: https://link.springer.com/chapter/10.1007%2F978-3-642-39360-0_29 

7. MIT TECHNOLOGY REVIEW. Astro, el robot de Amazon diseñado para colarse en casa y enamorarnos. October 11, 2021 [accessed 1/7/21]. Available at: https://www.technologyreview.es/s/13713/astro-el-robot-de-amazon-disenado-para-colarse-en-casa-y-enamorarnos 

8. FRONTIERS IN PSYCHOLOGY. Robot Authority in Human-Robot Teaming: Effects of Human-Likeness and Physical Embodiment on Compliance. May 31, 2021 [accessed 1/7/21]. Available at: https://www.frontiersin.org/articles/10.3389/fpsyg.2021.625713/full 

9. SCIENCE FOCUS. Future counsellor robots could comfort humans with a simple pat on the hand, new study suggests. May 5, 2021 [accessed 1/7/21]. Available at: https://www.sciencefocus.com/news/counsellor-robots-touch-study-future-technology/ 

10. IEEE XPLORE. Persuasive Robotics: The influence of robot gender on human behavior. 2009 [accessed 1/7/21]. Available at: https://ieeexplore.ieee.org/document/5354116/metrics#metrics 

11. IFRS. IFR Press Conference September 24, 2020 
Frankfurt. September 24, 2020 [accessed 1/7/21]. Available at: https://ifr.org/downloads/press2018/Presentation_WR_2020.pdf