Artificial General Intelligence: The growth of robots?
By Aditya Abeysinghe
Artificial Intelligence (AI) is the type of intelligence that current machines possess. It is based on learning methods such as deep learning and natural language processing. Using these methods machines could imitate humans and automate processes. However, AI-based processes cannot think similarly as humans when faced with unknown scenarios. Therefore, Artificial General Intelligence (AGI) has been used where cognition and other deep AI methods have been added to current AI-based methods.
What are the features?
A feature of AGI is the autonomous nature of problem solving. For example, systems which use autonomy to handle tasks use problem solving and creativity skills of AGI. Autonomous robots in vehicles and manufacturing plants are fine examples for such AGI-based autonomous systems. Techniques such as semi-supervised learning are used to experiment with data and unknown cases.
Another feature that can be expected from AGI is emotional intelligence. For example, humans love each other, share items and change attitudes. Robots using AI techniques rarely can express emotions and recognize others and respond to their emotions. If a human were to be replaced with AI, emotional intelligence is needed to change the artificial nature of current robots. Social skills are also important when speaking about emotional intelligence. Social skills such as handshakes, jokes, etiquettes etc. if present in robots enable them to be more human-like when mingling with people.
Logical reasoning, the ability to remember past events and arguments are also features of AGI. If the ability to take actions on their own is to be developed, reasoning on why something happens in that manner is important. This way AGI systems can improve thinking and be wiser in decisions and take measures to overcome threats towards them. Recalling past events also improve thinking and decision making as recalling what happened to past events could shape decisions on what happens in the present.
How AGI could change what is being already done by AI
AGI could improve efficiency of AI-based tasks due to the capability of AGI-based robots to handle tasks that cannot be handled by AI-based robots. One such area is motor robotics. For example, current AI uses specific set of logics to make robots perform operations. Robots which use these instructions can rarely perform other tasks which are not provided to them due to lack of thinking. Therefore, tasks such as logistics which require cognition and decision making is difficult to be handled by AI.
As AGI enables robots to behave as humans, they could be used in tasks which require continuous throughput. For example, robots could be used when human input is impossible such as in surveillance (24/7 support). Also, AGI could be used for analyzing existing data and then making decisions within lesser time to improve efficiency and reduce errors which take humans more effort.
The side effects
With AGI, robots have similar thinking, decision making and other human abilities which make them replacements for humans. Many have argued that they could replace humans not only in labor, but also in assistance. Therefore, the growth of deep AI needs to be in such a way that human race is not threatened by robots.
Image Courtesy: https://www.weforums.org