Metaphysics and Epistemology Group (Tuesday - Week 3, TT23)

The historically first and theoretically paradigm case of testimony is face to face, one to one telling.  But extended testimony comprises a range of other cases that diverge more or less from this paradigm: personal letters and emails; radio and TV broadcasts, newspapers and magazines, and all kinds of often anonymous sources of purported information on the internet. 

These still have a human or humans as their originating source. But now we have Alexa and Siri, the ‘polite young man’ who issues instructions from my phone GPS; and most recently ChatGPT.  I argue that this AI testimony is fake testimony: it fulfils the intended effect of its human creators through mimicking an intentional act of communication in a natural language,  but in fact is no such thing. I consider to what extent this affects the epistemology: does the difference in metaphysical kind entail a difference in epistemic kind - in how fake testimony can work to spread information?