Aleksandra Przegalińska, PhD: Our team is working on better understanding the nature of human fear of robots
What is a bot? It is a program the purpose of which is to replace a man. It specializes in conducting a dialogue and is often designed to mimic human behavior. In business bots are used in customer care to best respond to customers’ needs; they also prove handy in personalization of services provided to clients. They can be found for example on Facebook Messenger or in Wikipedia.
Botics means using knowledge about human behavior to model and simulate digital characters. Examples? Complex interfaces based on the analysis of natural language (NLP) and on conversational rules which improve man-machine relations. Botics uses the solutions of artificial intelligence (AI): expert systems, artificial neural networks, etc.
Chatbot (chatterbot) is a special type of bot. It is a computer program developed to converse in a natural language. Chatbots are also referred to as virtual advisors or lingubots.
Architecture of the bots
The most important element of a chatbot is the text analysis mechanism. It is a search engine which searches the knowledge base for information that appears the most relevant in a given situation and that is used to give meaningful answers which would reflect the reasoning of a human being.
The basic method of searching for answers consists in the use of two engines: the special one and the matrix one. The former compares the sentences provided by the user with the expressions and sentences stored in the database; the latter follows the same methodology to compare sets of single words.
Each instance of searching for answers consists of three stages:
- running the special engine,
- running the matrix engine (which compares the utterance of the user with the data in the database),
- using the so-called one-word file (the least effective strategy).
Works on creation of systems based on deep natural language processing, such as Duplex or most recent version of Cortana, are underway. In the future, also deep bots (mindbots) with developed cognitive functions (perception, associations, understanding of mental states of other agents) will be improved.
Uses of the bots
Customer care – automatic systems (assistants) to serve customers online (on platforms, in social media). They are used for example to provide clients with information about the offer and the enterprise, to create a positive image and to advise. Such solutions have already been implemented in Poland by ING Bank Śląski, InPost or IKEA. With such bots, companies can automate and personalize customer care.
Online games – in FPP games you can often encounter programs which target and shoot for players (aimbots). This kind of bots is used for training or when there is no access to the network. Bots are also present in MMORPG games where they are used to automate actions performed by a player.
Additionally, bots may be used in education (e.g. in learning foreign languages), therapies and care (telemedicine), entertainment or administration of intelligent houses.
Some Polish start-ups that develop bots
2040.io is a company from Cracow specializing in creation of modern IT solutions. By combining expertise with AI mechanisms, 2040.io wants to provide small and medium-sized enterprises with tools to accelerate their development. The company has been working on the Edward project, which is a chatbot that cooperates with sales agents by managing their agendas and reporting the results of their fieldwork.
Emplocity is a start-up that has created a bot used in the recruitment process. Emplobot also helps individual persons to find a job; you can talk with it via Facebook Messenger on the Emplocity start-up profile. The bot asks candidates questions about their CV, current profession, a city in which they would like to work, expected remuneration, etc.
AutomatedCheckin is a company from Białystok which helps travelers to check in at the airport. Its virtual assistant will ask you about your preferences connected with the flight and will perform a virtual check-in via Skype.
Fido.AI is an independent Polish research laboratory specializing in natural language processing technologies and AI. It has proven successful in the field of both commercial and research projects.
Inteliwise is an industry leading technology of semantic searching for quick replies; live chat and help center with FAQ, helpline and chat.
As in Poland there are many great programmers and data scientists, one might expect that, with the right amount of data, very good bot systems for administration, business, organizations and education could be created.
My research team is working on a scientific project connected with development of botics of the future. The core focus of the team is on the analysis of interactions between a man and a humanoid (social) robot and on comparing the features of that interaction with the features of communication between people. Robots are constructed here on the basis of expert systems.
Planned research follow the ongoing process of implementation of artificial intelligence in the areas of social interaction with people, placing particular emphasis on communication in the professional domain. In order to understand and to better explain a complicated relation between information technology and professional groups, the authors intend to evaluate how the abovementioned interactions influence social cognition of an individual.
We want to answer the following questions:
a) How and to what extent are the features connected with social intelligence present in robots? How do they manifest in interactions with human co-workers and clients?
b) How do social intelligence of robots and social cognition resulting from their interaction with human interlocutors influence the organization of work and cooperation?
The project addresses many important issues connected with the future of organizations developing systems that are more and more often based on information technology. How do such systems function in groups of people? To what extent can they do jobs that are usually done by people? Will we, as customers of such organizations, tolerate such systems and trust them or will we rather stay resilient? Where are the limits of applying artificial intelligence solutions to improve operations of an organization?
Communication with intelligent systems will be more intense and will require the increase of competences of both parties. It is important to understand how socialization of robots affects the organization of work and how the organizational culture emerges from such interactions. We assume that an emotional and physiological reaction will be the most intense in the case of contact between a man and a robot that is more similar to a human, that it will be less intense in the case of interaction with a “less human” robot, and that it will be the least intense with a chatbot.
What is the next step for humanoids?
At the end of 2017 Sophia, the first humanoid robot (a fembot, to be more precise), was granted citizenship. The machine is a citizen of Saudi Arabia. But even today it is not entirely clear what it actually means, what legal implications it has and what rights and obligations result from that for Sophia and her fellow citizens.
The decision might have been PR-related and the idea behind it might have been to lure as many scientists to the region as possible. However, one could also try to analyze it from the perspective of automation and robots becoming more and more commonplace in our life. Yet, although automation itself is an understandable tendency, constructing machines resembling a human being (and then making them citizens) remains unclear to the society.
Humanoid robots (and also social robots that do not resemble humans but have many anthropomorphic features) mimic men visually and functionally. Similarly to virtual bots (i.e. computer programs simulating human speech), they perform certain actions for men. Their function is often to imitate human behavior and to demonstrate how good they are at it.
However, humanoids are being used on a wider scale in business, they are increasingly more frequently “employed” in organizations, sales and customer service departments.
Both a typical humanoid and a bot are characterized by artificial intelligence, visual data processing and a face recognition system. They imitate human gestures and facial expressions, they are able to answer some questions and to have simple conversations regarding pre-defined topics.
Conceptually speaking, Sophia is similar to the ELIZA bot program, which was one of the first attempts to simulate human speech. On the other hand, it uses the Alphabet voice recognition technology and it has been designed in a manner allowing the robot to become more advanced as the time goes by.
The future of such robots in the western culture remains unknown, for instance due to the socalled Uncanny Valley effect referred to in professional publications on interactions between people and machines as a specific sense of “strangeness” accompanying a man interacting with the so-called virtual agents or robots.
However, our team is conducting works to better understand the nature of human fear of robots. Our current studies are to show to what extent interaction and cooperation between people and robots is possible and to what extent it makes sense, taking into account both innovations and never-ending problems with programming bots and robots.
Because of the fact that robotics and botics are stepping up their presence in our everyday life, it is necessary to fill gaps in research studies which, until now, have not taken into consideration a social and cognitive nature of interaction between people and humanoid robots.
We will also need to ask new philosophical questions about humanity boundaries, rights of robots or issues such as Searle’s “Chinese room” (which, by the way, also should be redefined in the context of development of machine learning). Such research has to be interdisciplinary and must be conducted by larger clusters or academic teams in consultation with the business community.
The article is based on a study by Aleksandra Przegalińska, PhD, entitled “Botics in Poland and around the world – research, development, implementation” (Warsaw 2018) prepared for the National Information Processing Institute.