The Tech Developers Group has been alarmed by the lack of understanding of the issues surrounding autonomy; the building, training and implementation of machine learning (ML) computer programme systems in particular. Such systems may be entrusted with all manner of data gathering, sensing, language and behavioural functions, and the range of issues that warrant consideration span from the basic to highly complex. We are concerned that the conversation amongst interested parties and delegates on the present and future capabilities of these technologies is either unrealistically utopian or dystopian, and largely fuelled by vested military or business interests rather than being based on sound technological understanding as applicable to the field use and law. Key players appear to lack a fundamental understanding of the basic structure of programming, training and the sequential processes which computers undergo when making “decisions”. Such understanding is crucial if ML is to be developed and implemented in a way that accords with humane values.
