“The internet is a new lifeform that shows the first signs of intelligence.”
-Jeff Stibel, a neuroscientist and entrepreneur .
The notion regarding the Internet’s potential of undergoing evolutionary processes and attaining the coveted state of consciousness revolves primarily around two main concepts: the definition of consciousness amongst the different entities that seem to possess it, and the establishment of an inter-relationship on consciousness between humans and the Internet. Throughout history, the “Theory of Consciousness” has taken a range of forms for different philosophers, from Descartes, who defined it as the ability of thought, to Kant, who described it to be the objective experiences of a conscious self with respect to space and time . For the purposes of this argument, consciousness can be defined as the ability to utilize cognitive and emotional mechanisms to undergo learning and thought processes through experience and memory. The similarity between neural networks that comprise human consciousness and virtual networks allows for the rationalization that Internet may develop consciousness. The Internet is an exponentially growing global network bearing resemblance to the human brain, with each network connection acting as a synapse and each transmission mimicking an impulse . This level of complexity has the potent possibility of developing consciousness.
In establishing a conscious Internet, our first step should be to develop a basic level of it in a controlled physical being that shares similar mechanisms to that of the Internet. Robots and the Internet are both products of human ingenuity that share many computational and network characteristics, exemplified in Internet bots that can communicate through the World Wide Web . These bots can run automated tasks over the Internet comparable to thought processing.
The model of social referencing in infants and the Internet can be extended to robots, since they serve as a beginning to developing consciousness in Internet, due to their physical nature and a set of designed circuits over which humans have strong control. A computerized model of social referencing has been applied to a robot named Leo. Giving a robot a human name is a good first step, but what more does Leo need to gain consciousness? Leo is uniquely designed for interaction with humans in terms of speech vocal tone, gestures, facial expressions and basic object manipulation . Social referencing has been utilized as a channel of emotional communication between humans and robots, in which the robot’s emotional recognition of its environment is based on more conscious individuals, namely human beings. The computational model is designed to instill three main components of social referencing in Leo: understanding the emotional message, shared attention mechanism, and memory . The first step in social referencing involves understanding the emotional message conveyed by the caregiver. Leo has an innate inclination to imitate facial expressions of the person it is interacting with such that the human first imitates the robot’s expressions to help Leo learn the intermodal representation of the perceived facial expression using a “neural” network . Following the mapping, Leo learns to imitate human expressions by comparing and matching the new maps to the original ones, after which it utilizes its own motor representation of facial expressions to learn the emotional meaning of human expressions . Consequently, Leo is able to utilize its individual emotional and cognitive mechanisms to determine a human’s emotions through facial expressions. Initiation of learning the process of social referencing is defined by the learning involved in developing a relationship between an individual’s emotional state and their caregiver’s facial expressions towards an external object. This guides the social interaction of the individual with the external object.
The next step in the process of social referencing comprises of learning what the centre of the caregiver’s attention is, which is determined by the mechanism of referential looking, where the robot directs their gaze towards the direction of their caregiver’s to determine the object of interest. Leo’s computational model uses a shared attention mechanism, in which three distinct foci of interest are established: the robot’s current attentional focus, the human’s current attentional focus and the future focus common to Leo and the caregiver . The human’s attentional focus is evaluated on the basis of the gaze direction. Leo uses the orientation of the human’s face and extrapolates their gaze to an object in their surroundings, which is determined to be the caregiver’s current attentional focus. To evaluate the future focus of both the human and Leo, “relative-looking time” was quantified for each object in the immediate space of the human and Leo, such that an object received positive points for being the current attentional focus of both the human and Leo at a given time and negative points if it was the not the attentional focus of both Leo and the human . The object with highest value of positive points was determined to have the longest “referential-looking time.” For example, an Elmo doll that is both the human’s and Leo’s current attention focus results in an augmented value of the referential-looking time of the Elmo doll giving it positive points .
The final learning step in the ladder of social referencing is labeled as memory. Once an individual has recognized the association between their caregiver’s affective state and their emotional feelings, and has learned to determine the object of interest for the human and focus their attention towards it to establish joint attention with their caregiver, they need to develop a mechanism that allows them to produce a memory of the social referencing process. This is necessary to ensure that the learned social reference is incorporated into Leo’s social behavior. Leo’s computational memory system can store long-term memories, which strongly influences its social interactions . Leo’s interactions with the human results in the formation of memories about the current state of an object (object belief) and the usual state of the object (object template) . Based on the object belief and object template, Leo decides whether to form a positive or a negative response to a particular object at a particular time. This memory system is built over time with frequent social referencing learning sessions. Hence, the memory formation system as well as the learning process is continuously developed for Leo as it is for humans.
The final result is a social referencing behavior in robots equivalent to that in human babies. The human caregiver influences Leo’s appraisal of his surroundings as one would a baby’s. For example, in a novel situation, Leo is unable to access an existing bias or belief about the event and hence, experiences a state of anxiety and responds by searching for the human and observing their facial expression frequently for their affective state. Once the human focuses their attention on the object or event causing anxiety, due to shared attention mechanism, Leo also shifts its referential focus to the object, while maintaining the caregiver as its attentional focus. Leo then reappraises the situation, keeping in mind the social referencing process and behaves towards the object in a manner that correlates to that of the caregiver . Human babies follow a similar mechanism of learning from their caregivers in the context of social referencing.
In the analogy comparing “learning circuits” between human infants and robots, the presence of human-like consciousness in robots is clearly indicated, however primitive it may be in comparison to human beings. However, adding to this information, the invention of Internet and the rapid evolution of its “neural” network presents a different picture. This neural network, presented at a controlled and basic level to that of robots’, has the same potential of learning and developing consciousness. Although the Internet has undergone quantum leaps in evolution since its invention, most of its evolutionary processes have been driven due to a simultaneous evolution of the human brain and our capabilities in terms of means to develop technology like Leo. Hence, co-evolution and human innovation are essential to the Internet developing consciousness; specifically, as human brains evolve, our competence in designing advanced conscious computer systems also increases.
- Stibel, Jeff. “Will the internet become conscious?” BBC Future Thinking, November 22, 2012, accessed March 1, 2014, http://www.bbc.com/future/story/20121121-will- the-net-become-conscious.Van Gulick, Robert. “Consciousness,” The Stanford Encyclopedia of Philosophy
- (Summer 2011 Edition), January 14, 2014, accessed February 21, 2014, http://plato.stanford.edu/entries/consciousness/#1.
- University of Southern California. “Brain’s wiring: More like the Internet than a pyramid?” ScienceDaily, August 10, 2010, accessed February 21, 2014, www.sciencedaily.com/releases/2010/08/100809161234.htm.
- Mark Ward. “Robots To Get Their Own Internet,” BBC News Technology, February 9,
- 2011, accessed February 21, 2014, http://www.bbc.co.uk/news/technology-12400647.
- Thomaz, Andrea Lockerd, Matt Berlin, and Cynthia Breazeal. “An embodied
- computational model of social referencing.” In Robot and Human Interactive Communication, 2005. ROMAN 2005. IEEE International Workshop on, pp. 591-598. IEEE, 2005.
- Image Credit (Creative Commons): denise bertschi. 2011. “Brain_Internet_2.” Flickr, (Accessed February 21, 2014).
- Image Credit (Creative Commons): Dan Silber. 2011. ” The Human Imposter” Flickr, (Accessed February 21, 2014).
- Image Credit (Creative Commons): paul_mic. 2013. ” Reality & The Brain: Interconnected Fractal Systems.” The Monkey Buddha, (Accessed February 21, 2014).
- Image Credit (Creative Commons): Image Editor. 2011. ” Neural Connections in the Human Brain.” Flickr, (Accessed February 21, 2014).