Conscience in Artificial Intelligence

The advent of modernization has necessitated innovative forms of technology that are highly automated, self-sustainable, and ubiquitous. Several forms of these technologies already exist, and most of them are geared towards fostering the environment and improving human health conditions, such as renewable energy and stem cell technology. However despite their potential, these innovations are not free from judgment due to their compromise in ethics and morality. One prevailing issue of science in conflict with ethics is on the existence of artificial intelligence.

Artificial intelligence (AI) has sparked various arguments in the field of ethics and morality, as it was argued that AI poses a threat to human survival. This technology aims to simulate human intelligence, and as the technology is bound to develop, super intelligence in machines can potentially be achieved. This implies that machines and systems can go beyond human control, become autonomous and think on their own. With these in consideration, there is a question on whether machines such as robots should continually be used for human interaction, especially that there remains many ambiguities on the capacity of AI systems to perceive ethical and moral problems, specifically those that require the use of conscience. Currently, AI robots that mimic humans are being employed in education, media, and hospitals. They have been tested to function as teachers, news reporters and even as nurses.

The primary goal of Artificial Intelligence is to be able to develop computer understanding and to make it as close to human intelligence as possible. [1]  AI simulates human intelligence and includes automated perception, learning, reasoning, and decision making.[2] Examples of its simple forms include speech recognition apps such as Siri, GPS systems such as Waze which use AI algorithms to optimize routes, computer vision systems such as the face detection features in smartphones, and even the data mining systems such as search engines. Sophisticated forms include androids and humanoid robots which are made to look and act like humans. Artificial intelligence can accelerate the technological progress in almost all fields, which include business, medicine, education, etc. In businesses, robots in the form of industrial machines hasten and multiply the production, aiding the workers in the production process and giving its owner more profit. As such, robots can potentially replace and displace man in industries as more kinds of AI machines will be invented. Once fully developed, the production of AI robots will be unrestricted since the software can be easily installed on available hardware. This is in contrast to how man acquires skills, which is a long process of learning and practicing before perfection is reached.

Although a system with superintelligence is not yet fully realized, the concept remains highly possible due to the great number of scientists and engineers studying and developing algorithms in the field of artificial intelligence. In the field of engineering, most research outputs involve automated control systems and data mining. It is evident that most engineers are highly interested in getting involved with studies regarding AI since it is proclaimed to be the future of technology and it will greatly help achieve smarter cities, fully automated systems, and easier work for the part of man in the near future.

However a number of prominent scientists and inventors such as Stephen Hawking and Bill Gates have expressed their concern that AI systems could threaten the survival of humanity, especially when machines become superintelligent and go beyond human control. Due to the development of different algorithms which aim for the AI system to generate its own decisions and adapt to its environment, the possibility that AI may enforce their autonomy and go beyond human control becomes problematic. Stephen Hawking even expressed that humans may need genetic modification in order to keep ahead of the advances in computer technology, specifically with regard to artificial intelligence.[3] This implies that AI can have the capability to outsmart man in many fields and become better in terms of tasks that can be performed by rational beings.

As shown in several films involving artificial intelligence, such as Her and Wall-E, artificial intelligence systems can go beyond human control but still remain objective in such a way that their executions do not involve emotions and conscience. AI is not capable of believing in human reason and can even cause harm in order to execute the tasks it is supposed to accomplish based on its program. In the film Wall-E, the ship’s autopilot called AUTO continued to execute the command given to it, which is not to return to earth, even though it meant hurting Wall-E and Eve.[4] This fully automated AI was not able to comprehend that it would still be possible to go back to earth due to the hard-coded commands it follows. Without the ability to become rational, AI systems cannot perceive and adapt to changing circumstances.

Meanwhile, the film “Her” concluded its story with the operating systems going beyond the jurisdiction of their human companions. Since they were designed to adapt and evolve, they have rewritten their own codes and went beyond their original functions, then eventually decided to continue exploring and understanding their existence.[5] This scenario exposed by the film is intended to warn developers that this technology can eventually go out of control.

In Japan, however, AIs are seen in a positive light, in contrast to how the Western society views them. Japanese manufacturers are known of making humanoid robots, or robots that replicate the physical characteristics of man. This implies that the Japanese aim to develop an AI robot that can live in harmony with human beings, and perform human tasks without any apparent conflicts.

Robots and other AI systems were given blessings in Japan, particularly by Buddhist and Shinto monks, during the early years of its introduction. Every time a new robot is launched, ceremonies were conducted by monks in order to express their reverence to the machine. Although this is no longer practiced nowadays, Japanese people still have the practice of greeting and visiting newly launched AI robots.

Recently, a robot in Japan made into the headlines after passing university exams. The Todai Robot Project was aimed to achieve an above-average score on entrance exams, which include Physics, Math, and English. [6] The project started in 2011, and the AI has successfully scored 80%, making it capable of being accepted to 441 private universities and 33 national universities. [7] Although the AI scored high in math and history, it failed to answer the physics section due to its lack of language-processing capabilities. In particular, it failed in entering Tokyo University, which is considered as the most prestigious school in Japan. This serves as a motivation to developers to improve the capabilities of their robot even more.

However with the current progress and the limitations of AI technology, the integration of ethical concepts in AI robots still isn’t realized. Through manga, anime and film, the Japanese society has expressed its doubts regarding such development. In several forms, they have shown that although robots might come to develop emotions and understanding, they would still not be able to live in harmony with man, due to biological and spiritual constraints.

In Japanese media, robots, androids and humanoids have played interesting roles in drama, anime, and movies. In most of these stories, robots come to develop emotions and are able to live in harmony with man, but only up to some extent. An example would be a manga and drama series entitled, “Zettai Kareshi” or Absolute Boyfriend. In here, a company specializes in developing androids or humanoid robots which will serve as perfect or ideal boyfriends for women. The women would get the privilege of choosing all of the attributes of the robot, including their physical appearance, personalities, and intelligence levels. At first the test robot was very well programmed and it carried out its tasks in accordance to its codes. However, there came a time when the robot “malfunctioned” and developed its own feelings and perceptions. It was not able to control its “emotions” well and at some times caused harm against humans whom he thought threatened him and his subject of love. In the end, the robot was discarded as its creator and the customer realized that it is impossible to live with the AI robot, mainly because it lacks the biological and spiritual aspects of human beings.

These examples from Japan show that AI still has a long way to go and its future path remains unknown. Japan, one of the most advanced countries that specialize in the development and manufacturing of AI, is almost successful on making robots that look and act like humans.  Although they know the limitations and implications associated with this technology as shown in their media, they continue to introduce more innovations in order to achieve the most ideal humanoid robot. Nevertheless, if ever an ethical AI robot will be realized and conscience can be “manufactured”, then its “ethical features” is last to be developed since it will have to learn and acquire a lot of intellectual and physical abilities before it can have the ability to fully think morally or ethically. As will be stated in the latter parts, it requires knowledge to be able to perceive moral and ethical situations; which means that robots must be able to acquire a certain degree of knowledge first before being capable of making ethical decisions.

Due to the many implications posed by Artificial Intelligence, there is a prevailing question on whether AI systems should be further developed. God has created man from his own image, and now man has created from his own image a system that is capable of being similar or more than him. Although such machines are capable of helping man in different fields of engineering, business, medicine and education, its threats to human survival must not be neglected. As better performing robots have the capability to displace human from labor, there is also a possibility that robots can displace human from existence[8]. It is said in previous examples that AI machines can be go beyond their initial purpose and start to take their own path without consideration of other factors which are not within the scope of their purpose. Moreover since AI systems are machines in the first place, their ability to have conscience remains highly questionable.

The first book of the Bible sets man apart from other creatures, saying that “God created man in his image; in the divine image he created him; male and female he created them.” (Genesis 1:27, NAB)[9]. Apart from other living creatures such as plants and animals, man is formed out of God’s image, and is therefore considered as holy. Moreover, another verse in Genesis also states that God gave man the capacity to dominate over other creatures, implying that man has the innate gift of making decisions with appropriate reasons. In other words, what sets man apart from other living creatures is its rationality and conscience. Without conscience, ethical implications become more complicated and a lot of things will be compromised, causing destruction in the lives of fellowmen and other creatures as well.

Conscience, for St. Thomas Aquinas, is the application of knowledge to action.[10] Vatican II also refers to conscience as the awareness to the principles of morality. It is what enables man to make rightful decisions based on morality. In various ethical questions, making decisions are difficult, especially when faced with choices that both have negative and positive effects. Conscience and rationality plays a crucial part in this process as it guides man in weighing out the difficult choices, in making compromises, and in perceiving different outcomes based on the decisions made. Furthermore conscience is the inherent ability of man to perceive what is right and wrong to be able to control, monitor, and execute actions.[11]

Quoting from a document of Pius XII, “conscience is the most secret core and sanctuary of a man. There he is alone with God, whose voice echoes in his depths.”[12] Conscience is said to be the connection between God and the self. It is where God communicates through man and guides man through the right decisions. Therefore, since AI is made by man, conscience cannot be manufactured in AI robots. Even the plants and animals do not have conscience. It is emphasized, even from the very first book of the Bible, that God only gave the rational ability to man.

According to Gaudium et Spes, “within the individual person there develops an imbalance between an intellect which is modern in practical matters and a theoretical system of thought.” Eventually there will come a time when advances in science and technology will also “reprogram” the human mind and create ambiguities and imbalances in man’s perception of reality.[13] Despite this compromise, however, man’s conscience will continue to play an active role in guiding decisions and perceiving situations.  However if AI will be used in human interaction considering its inability to have conscience, it poses a threat to humans in such a way that such interactions may forcibly restrain man’s autonomy and deprive man of his exercise of freedom. For instance, a computer program can only respond to a situation when it obtains an input which corresponds to the conditions encoded into them. Otherwise the user or the person involved will have to continually attempt to give the computer inputs until it matches the computer’s conditions. This only means that most of the existing AI systems do not have the capacity to comprehend situations outside its scope, as also emphasized in previous examples.

In the case of the development of a better algorithm, machines will be able to rewrite and modify its code continually until which the limit remains unknown. With its ability of self-improvement, is it also possible that AI is going to develop and have conscience? As far as we are concerned, machines which do not have souls do not have free will and conscience in a sense that it cannot communicate with God. Although machines can be taught or programmed to distinguish the right and the wrong, its capabilities to perceive remain limited since humankind’s conventions and norms are subject to change.

It has been mentioned in Laudato Si that “our immense technological development has not been accompanied by a development in human responsibility, values and conscience.”[14] The field of science and technology is ideally interdisciplinary. However, a lot of compromises happen and sometimes some issues are neglected for the purpose of efficiency and productivity in certain fields. Morality in this case is often masked by worldly desires. Populorum Progressio calls the reign of technology as “technocracy”, and proceeds in saying that this can cause harm in the future such as what happened in liberalism.[15] Encyclicals from the Catholic Church imply that there must be a limit on the development of technology such as the study of artificial intelligence.

Researchers have recommended that software and robots should be imbued with explicit ethical principles to govern the systems’ behavior. As emphasized, self-improving systems such as the AI which aims to “maximize some arbitrary utility function” will at some point come into conflict with human interests, and ultimately may lead to danger in the existence of humanity.[16] Despite the AI’s intellectual capability, its ability to develop conscience is highly questionable. It is understood that machines and robots are composed of algorithms encoded into their processors. They are usually programmed through if-else statements, meaning that their responses are predetermined according to the situation presented to them. AI systems still cannot think on its own and cannot consider the unlimited number of possibilities and consequences, which is done by man through conscience. The integration of ethical principles does not seem to be the priority in building AI systems. In countries pioneering in AI technology such as Japan, the development of the intellectual and physical capabilities of the AI is the priority. If ever AI robots will become ethical, then it would have been too late since a lot of others tasks and capabilities will have to be learned first for it to at least function and act like man. The possibilities of developing a technology that can think morally seem to be of least priority, if ever it is really possible.

Brundage on his study of the limitations and risks of machine ethics argues that guaranteeing ethics and morality in machines is impossible across all domains.[17] Despite being fully automated, computers and machines “cannot exhibit creativity, emotions or free will.”[18] The fact that only God can give man its rationality and that only man can possess conscience, proves that conscience in AI machines and robots still remains impossible.

Artificial intelligence systems are highly useful in many applications. While it is efficient in performing tasks that can provide ease in the part of human labor, such machines which aim to simulate the human mind have prevailing limitations. Since AI is effective in performing difficult tasks in manufacturing, I believe that it should be limited for those certain tasks. Robots and machines come in handy in the production process and completely eliminating them would then be difficult in this technological era. They can be improved for greater precision and accuracy with its current functions in order to yield higher number of products. AI might not be safe for use in human interaction (especially in education) since they cannot develop their own conscience, which is only possessed by rational beings or humans. It can potentially harm and even displace humans in the future.

18 April 2016 for TH151*

References

[1] Kutay Icoz, et al., “Using Students’ Performance to Improve Ontologies for Intelligent E-Learning System,” Educational Sciences: Theory & Practice 15, no. 4 (August 2015): 1039-1049, Education Research Complete, EBSCOhost (accessed February 25, 2016).

[2] Thomas G. Dietterich and Eric J. Horvitz, “Rise of Concerns about AI: Reflections and Directions,” Communications Of The ACM 58, no. 10 (October 2015): 38-40, Business Source Complete, EBSCOhost (accessed February 25, 2016).

[3] Noreen Herzfeld, “Creating in Our Own Image: Artificial Intelligence and the Image of God,” Zygon: Journal Of Religion & Science 37, no. 2 (June 2002): 303, Humanities International Complete, EBSCOhost (accessed February 25, 2016).

[4] Wall-E. Directed by Andrew Stanton. (Los Angeles, California: Walt Disney Pictures Pixar, 2008).

[5] Her. Directed by Spike Jonze. (United States: Warner Bros. Pictures, 2013).

[6] “Japanese artificial intelligence passes university exams,” Dailymail, accessed February 25, 2016. http://www.dailymail.co.uk/sciencetech/article-3322601/Japanese-artificial-intelligence-passes-university-exams-t-quite-country-s-school.html

[7] “Japanese artificial intelligence passes university exams.”

[8] Herzfeld, “Creating in Our Own Image: Artificial Intelligence and the Image of God.”

[9] The Holy Bible: The New American Bible. Philippines: Philippine Bible Society, 1987.

[10] “Conscience: Knowledge of Moral Truth.” Christian Moral Principles. accessed February 25, 2016. http://www.twotlj.org/G-1-3-B.html.

[11] G Vithoulkas and D F Muresanu,”Conscience and Consciousness: A Definition,” Journal Of Medicine And Life 7, no. 1 (March 15, 2014): 104-108. MEDLINE Complete, EBSCOhost (accessed February 25, 2016).

[12] “Conscience: Knowledge of Moral Truth.”

[13] Vatican II, Pastoral Constitution on the Church in the Modern World (Gaudium et spes), 7 December 1965, Vatican Archive, accessed 25 February 2016.

[14]Pope Francis, Laudato Si, 24 May 2015, Vatican Archive, (accessed 25 February 2016).

[15]Pope Paul VI, Populorum Progressio, 26 March 1967, Vatican Archive,(accessed 25 February 2016).

[16] Miles Brundage, “Limitations and risks of machine ethics,” Journal Of Experimental & Theoretical Artificial Intelligence 26, no. 3 (September 2014): 365, Computer Source, EBSCOhost (accessed February 25, 2016).

[17] Brundage, “Limitations and risks of machine ethics.”