Computational Consciousness is the new Artificial Intelligence

The acronym 'AI' prioritises the intelligence (ie problem solving ability) of information processing machinery. Since the 1950's when this acronym first appeared, the industry focus on intelligence guided improvements in computer science, which centred upon the manipulation (eg target element search, element insertion) of data structures which were optimal representations of the problem domain. The implication was that the main task that brains performed was the solution to (complex but well-formed) problems, such as playing chess, or participating in a natural language dialogue at a teletypewriter (as in the Turing Test). If other aspects of cognition were deemed valuable, it was only because they were constituents of 'clever' or 'smart' machine behaviour.

This analysis is not intended as a critique- indeed, the leaps in both the performance of computers, as well as their low-cost and ubiquitous availability is a direct consequence of this academic focus on intelligence. Intelligent machines are, within this conceptual framework, fast machines with large memories, because intelligence is equated to the search of large data representations for solution states to problem spaces. This reveals the true importance of Moore's Law, which predicts the doubling of computational performance every 1.5 years. Without such serious improvements in 'horsepower' occurring on a regular basis, expected gains in 'intelligence' are unachievable. This doubling of performance hides some technological advances which are often an order of magnitude. Consider typical PC memory capacity, which increased 1000 fold from 20MB in the mid 1980's to 200GB in the mid 2000's.

Recently, research efforts seem to be directed at other components of cognition, such as machine analogs of consciousness and emotionality. The commonly stated reason for this change in focus is that some major AI research topics such as natural language and natural scene understanding have stalled. In reality, this is not the case. AI of all types, especially 'deep' machine learning techniques, are now big business. For example, facial recognition software is more than intelligent enough to become an indispensable tool of many of the world's secret security services, and lately, public police forces too.

Just as the original focus on intelligence gave rise to the ubiquitous near-zero-cost PC (2GB RAM Arduino for $50), the fruits of recent research into consciousness and emotionality will no doubt include safer autopilot software for airplane and driverless automobiles, not to mention more intuitive, and individuated (ie personalised) graphical user interfaces, especially for mobile telephony apps.