With the 91st Academy Awards coming up on the 24th February it is a timely prompt to consider both the treatment of Machine Learning (ML) and Artificial Intelligence (AI) as subjects of movies, as well as the increasing influence ML and AI have on their consumption.
Cinema as an art form has been driven by technology since its inception. The development of both cameras and media on which to record or play films has had a symbiotic role with the subject matter of movies.
As subjects of films, both ML and AI have been present, albeit in nascent form, since Fritz Lang’s Metropolis in 1927. Almost all histories of movies featuring ML start there and, along with AI, it has been a recurring theme. Noteworthy examples include; 2001: A Space Odyssey, War Games, I Robot, AI, Her, Ex Machina and Lucy.
There is a tendency for ML in particular to be portrayed in a negative, often malevolent light; as having an agenda of its own, contrary to that of the human protagonists. There are exceptions: in War Games from 1983, the computer WOPR eventually learns the futility of nuclear war by repeatedly playing every possible version of noughts and crosses/ tic-tac-toe.
A much more nuanced approach is taken with the AI computer HAL 9000 in 2001: A Space Odyssey. Here the computer possesses cognitive circuits, which develop faults, impairing its ability to function effectively when presented with contradictory mission imperatives.
For arguably the most realistic depiction of data science it is best to consider movies based on factual events. The use of data features heavily in both Moneyball and The Imitation Game.
Increasingly ML is being utilised to analyse all aspects of our consumption of movies. From day to day monitoring of our viewing preferences on YouTube in order to tailor suggestions or advertisements, to algorithmic analysis of thousands of films’ emotional arcs to categorise and codify each one as well as determine the likely success of a specific film or even to use ML data analysis to predict likely winners of Academy Awards.
There is progress ongoing in all of these facets of ML and AI in their application to usefully predict or influence our choices and preferences. There may come a time when, not unlike the Black Mirror episode Bandersnatch where the viewer makes choices that affect the plot and outcome of the story, ML algorithms will determine what version of a movie will be most appropriate to each individual viewer.