Read on Twitter
As part of my OLPOAI (Officially Licensed Philosopher of AI) duties, I'm going to start a thread of recent papers I like in philosophy of AI.
First up is @KathleenACreel's excellent "Transparency in Complex Computational Systems" https://philpapers.org/rec/CRETIC
First up is @KathleenACreel's excellent "Transparency in Complex Computational Systems" https://philpapers.org/rec/CRETIC
In a cutting edge paper on opaque algorithms, Kathleen actually utilizes an old-school favorite move of the analytic philosopher: making a distinction!
No, but seriously, everyone working on opacity and explanation in AI should use the sophisticated framework she presents here
No, but seriously, everyone working on opacity and explanation in AI should use the sophisticated framework she presents here
In case anyone wants to spend a cozy Sunday morning reading papers on the philosophy of AI (and who doesn't?!), next up on my list is @cameronjbuckner's "Understanding adversarial examples requires a theory of artifacts for deep learning."
https://www.nature.com/articles/s42256-020-00266-y.epdf?sharing_token=tf2AjxdyxaCpRZDh-bgry9RgN0jAjWel9jnR3ZoTv0PIZwnfXXIXpRQcD8xNIztDbtfVqTyBDH9X-OOXzY7eZ-IxEmR2Hp4l-jkk7SsFNfq5YF28S1ed3ku5PhVGAWq8CU6EhCQXTsZtqXhxceo1GCaH5BpK4cqsUrQ8Mbf4UeU%3D
https://www.nature.com/articles/s42256-020-00266-y.epdf?sharing_token=tf2AjxdyxaCpRZDh-bgry9RgN0jAjWel9jnR3ZoTv0PIZwnfXXIXpRQcD8xNIztDbtfVqTyBDH9X-OOXzY7eZ-IxEmR2Hp4l-jkk7SsFNfq5YF28S1ed3ku5PhVGAWq8CU6EhCQXTsZtqXhxceo1GCaH5BpK4cqsUrQ8Mbf4UeU%3D
I'm biased, because I gave comments on this paper at this year's SSPP. But Cameron's work in this area is both so technically proficient and so philosophically interesting that I think it should be on basically everyone's radar
Also, if any philosopher wants an introduction to the deep learning technologies that drive much contemporary discussion of AI, Cameron has that covered too: https://philpapers.org/rec/BUCDLA
My next recommendation is probably known to basically anyone who does any work in this area: @ShannonVallor's authoritative "Technology and the Virtues" https://global.oup.com/academic/product/technology-and-the-virtues-9780190498511?cc=us&lang=en&
I am not myself a virtue ethicist, but I cannot imagine a more thorough and illuminating account of a virtue-ethical response to technological advancement than Shannon's book. If this relatively new field has any "foundational" texts, this is definitely one.
Continuing this thread, the intersection of AI and comparative psychology presents us with a ton of interesting questions, like: can AI be creative? @MartaHalina has a fascinating new paper on this subject:
https://onlinelibrary.wiley.com/doi/pdf/10.1111/mila.12321
https://onlinelibrary.wiley.com/doi/pdf/10.1111/mila.12321
Her answer is an intermediate position. She argues there is one aspect of creativity (mental scenario building) that AlphaGo (for instance) can accomplish via Monte Carlo tree searches. But other aspects of creativity, like domain generality, remain beyond its reach.
