The two main philosophical questions with which artificial intelligence (AI) has been traditionally concerned are (1) 'Could a machine think?’ and (2) 'Are we (humans) thinking machines?’ (see Walmsley 2012). Recent work in AI has continued to seek answers to these questions either by building technological tools to perform activities and accomplish tasks that human minds can do, or by helping us to understand the processes and mechanisms involved in human cognition. In addition, recent work in AI attempts to go beyond the human case to raise questions about the cognitive capacities of actual and potential systems that are more powerful than the human mind, and to understand the consequences—and the risks—of the underlying technological developments.
Recent extensions of previous work in AI include the development of ‘Deep Learning’ algorithms to enable artificial systems to learn from complex data with much less intervention and supervision by the human programmer. This is partly inspired by success in neuroscientific research on the structure and function of the human brain (and developments in connectionist and neural network AI), and has a wide range of technological applications and psychological implications. Such work goes hand-in-hand with work in brain emulation and the fine-grained replication of neurological structure in non-biological substrates.
In addition to building on earlier successes, recent work has begun to address related questions about the consequences – and risks – of the development of AI technologies. First, there is a significant (and increasingly nuanced) debate about the possibility of (and the timeframe for) the development of ‘superintelligent’ AI (AI systems that exceed the cognitive capacities of humans): a hypothetical point in the future that has come to be known as ‘the singularity’. Second, recent work in AI has seen a growing appreciation of the risks of both current, and hypothetical future, AI.
As a consequence, recent work in AI continues to address—and shed light on—many familiar philosophical questions. In addition to the general questions of whether machines could think, or whether the human mind could be understood in mechanical terms, there are also several specific questions that touch on other areas of philosophy, such as: would the development of AI show that physicalism, or functionalism, about the mind is correct? How should we understand the identity of a person over time? How should we understand the relationship between phenomenal consciousness and the brain? What is the relationship between values, motivation and behaviour?
Movement, activity and action: the role of knowledge in the perception of motion
