Does Artificial Intelligence Have a Consciousness?

What if the robots in films Her and Ex Machina were real? Well, there is such a robot and her name is Sophia, as shown in Figure 1. Loved for her exceptional human interaction, Sophia was granted citizenship in Saudi-Arabia in 2017. Although Sophia has numerous human-like qualities, she is merely an artificially intelligent robot. Artificial intelligence, as defined by Illinois Technical Institute Professor Elisabeth Hildt, is intelligence demonstrated by robots. Technology has developed to the point where robots are behaving as humans, and a common question about these technological advancements is if artificially intelligent robots can possess a consciousness. Can these robots show true emotion? Do robots have a consciousness?

Figure 1. Robot Sophia Source: Robots IEEE

Human vs. Robot

It has been 3 years since robot Sophia was granted citizenship, and it has been even longer since she was first created. There have been major advancements and much has changed in the engineering field since Sophia was first created. Questions regarding robot consciousness have been raised, creating diverse views. Associate professor and PhD Eugene Piletsky explains humans must understand their own consciousness in order to design some sort of equivalent consciousness into robots. In his article, Dr. Dagene Song, a quantum theory expert, correlates language and culture to consciousness. Because robots are created, they are not somehow, "programmed with culture." Most engineers relate consciousness with emotions and empathy. Although robots can be programmed to show empathy or express emotions, PhD Harry Haladjian and colleague Carlos Montemayor further explain how these emotions are created, this signifies, well, artificial emotions. These emotions would solely be simulations. Human moral reasoning is based on experiencing empathy and emotions, however, robots do not necessarily experience this. Although robots can show great levels of intelligence, simulated emotion is not considered to truly be emotion. Dr. Piletsky also advocates this argument because he believes we humans are born with motivations and emotions, and even if robots were to gain some form of self-awareness, this does not automatically implicate possession of emotions or desires.

Designing a Conscious Robot

Because current robots do not exactly fit human moral reasoning due to lack of experience, University of Oxford researcher Soumya Banerjee comes up with a framework, as depicted in Figure 2, on how to design a compassionate and conscious robot.

Figure 2. Framework for conscious robot Source: Soumya Banerjee

The design begins with the core engine where concepts are formulated, and data is analyzed. The robot then has several modules programmed at the core. The empathy module is how the robot can understand others. Next is the consciousness module, where the information processing occurs. The consciousness module will allow robots to communicate with other artificially intelligent machines. Banerjee even includes a dream module, and this acts as a "sandbox environment," a method of testing software in an isolated environment using virtual servers. The dream module will basically turn on all of the other modules, but the robot will not inhibit a physical response. The dream module is closely connected with the experiment module, where most of the experimenting occurs. The last couple of prime modules are the explainability module, and the memory module. The memory module will allow the robot to accumulate experience, to then further express empathy and emotions. With the memory module, the robots can also learn from past mistakes, and adapt to different situations. Although this framework may seem convincing, there are a great deal of limitations.

Limitations and Objections

Psychologist Ricardo Manzotti and lab specialist Antonio Chella suggest in their article, in order to design a conscious robot with a first person view, a computational approach must be taken. Current robots, however, are limited to a third-person perspective, and a computational approach adds to the robot's intelligence, but not so much its consciousness. Artificially intelligent robots are in need of a past to be able to show signs of consciousness through exhibiting emotions. Dr. Dagene Song suggests a physical robot or entity implies a future, but a consciousness implies a past. Currently, there are various non-computable programs that cannot be duplicated in self-aware conscious robots. A computational approach may suggest a robot is intelligent, but human intelligence is based on a sense of purpose, and this purpose relies on various factors such as biology, history, and culture. Although these robots will have the knowledge of a wise person, their motor skills are of a baby. Designing a robot that does not need human assistance is not an easy task. For example, current robots would need help crossing a road, but they could win a game of chess with ease. Motivation and desires are also a vital factor that play a role in robot consciousness. Empathy is a motivation, and robots cannot show empathy, because they lack motivation. Robots can, however, read and learn how to show empathy, but they cannot genuinely empathize or understand social interactions. This ties in with motivational halting. To explain this phenomenon, allow me to provide an example. If someone, say your parents, were to aggravate you due to opposing views on a subject, you would be able to control yourself and avoid saying anything spiteful, because you do not wish to offend your parents. If a robot were to be in the same situation, the AI would not know when to stop arguing. Algorithms can produce simulations that make robots appear as humans, but feelings or emotions are part of our neural system, something a robot does not possess and cannot be replicated.

Robotic Consciousness Examples

Perhaps the greatest example of robotic self-awareness is their ability to walk independently. This physical independence adds to ability to have a consciousness. Although robots lack motivation to show emotions, they will instead go into what is called a virtual nirvana, almost like a virtual reality state, as Banerjee explains. This is significant because the virtual nirvana state is equivalent of the robot's thinking or imagining process. The creator of the machine has the power to decide what they would like to program into the unconscious robot. Once this is done, the robot, when "conscious" or turned on, will reflect the traits the creator programs onto them. Lastly, tests have been ran to get two robots to communicate with each other, rather than a robot only communicating with humans. When this was done, the two robots created their own code of communication, and the research had to be stopped.

Conclusion

So, do robots have a consciousness? Current robots are not designed or capable to have a conscious, or, in clearer terms, show signs of having a consciousness. Research has allowed engineers to come up with what conscious robot designs should be composed of. Further research must be done to implement these research findings. Robots can only mimic humans, but they are not living and culturally knowledgeable beings.