Robot Brains: Circuits and Systems for Conscious Machines

11

Machine consciousness

11.1 CONSCIOUSNESS IN THE MACHINE

The ultimate goal of machine cognition research is to develop autonomous machines, robots and systems that know and understand what they are doing, and are able to plan, adjust and optimize their behaviour in relation to their given tasks in changing environments. A system that succeeds here will most probably appear as a conscious entity. Here the issues of the apparent properties of the human consciousness and those of the proposed machines are discussed.

At first, it is useful to make the difference between the concepts of the contents of consciousness (such as the immediate percepts and memories) and the mechanism that makes the system aware of the contents of consciousness. The machine consciousness research is mainly interested in the mechanism of consciousness, while the contents of the machine's consciousness, as interesting as that will be, are here a secondary issue. Additional confusion may arise from the fact that the phrases ‘to be conscious of something’ and ‘to be aware of something’ are often also used in the meaning ‘to know’ or ‘to have the knowledge of something’.

A conscious agent is aware of its existence in a meaningfully perceived world. Therefore, in order to understand the world the agent must deal with meanings. This is a fundamental requirement. If a machine that is claimed to be conscious does not process meanings then the machine is not conscious. Real-world meanings can be extracted by suitable ...