Eindhoven – The project explores how we can train machine-learning algorithms to become more emotional.
In her project Still Life, Vera van der Burg trains a speculative algorithm with a highly subjective set of rules to emulate a human understanding of objects. These algorithms ambiguously learn from patterns in large data sets and interpret a still life photograph through emotional values such as love, jealousy and seduction.
With AI becoming more pervasive in changing human behaviour, consumers are becoming more inquisitive about the ethics powering our data systems. In a similar vein to Trevor Paglen’s From ‘Apple’ to ‘Anomaly’ (Pictures and Labels) exhibition, Van der Burg explores how machine-learning systems are biased by the programmers behind them. In a statement about how self-learning algorithms are often a reflection of their subjective values, the object-recognition algorithm in Still Life is a reflection of Van der Burg and her tastes.
To learn more about algorithmic ethics and how to use design cues to harness machine-learning systems, read our Morality Recoded macrotrend.