On 3/10/22 18:44, Roger Clarke wrote:
Discussion here:
http://www.rogerclarke.com/EC/AII.html#TAA (2019)
Just as a comment, Roger's "discussion" states:
Of particular concern are assertions that empirical correlation
unguided by theory is enough, and that rational explanation is a
luxury that the world needs to learn to live without. These cavalier
claims are published not only by excitable journalists but also by
influential academics (Anderson 2008, LaValle et al. 2011,
Mayer-Schoenbeger & Cukier 2013).
This position has been advocated previously on Link in relation to AI
and self-driving cars. And while lane-keeping and automatic-braking
technology isn't usually blessed with the holy oil of "AI", once enabled
they are clearly, to use Roger's analysis, Level-7 Decision Systems in
the heat of an emergency.
The "cavalier claims" above reflects an old argument in the philosophy
of brain, mind & freewill which holds that a person's actions can only
be considered in terms of their external, observable effects, and never
in terms of an actor's "intentionality" or "state of mind" because these
are hypothetical, unobservable, internal states. Thus "anger" is
interpreted as a propensity to act in a certain way, not as a state of mind.
One of the main proponents of this view was the British Philosopher
Gilbert Ryle.
The argument that self-driving cars AI-controlled are OK if they result
in some nett reduction of "harm" (take your pick as to how we might
assess that!) seems to me to be quite untenable. How such an
interpretation might fit into all human legal frameworks given that the
notion of personal responsibility goes out the window is surely a
mystery, and I'd argue the whole behaviourist argument is flawed anyway.
We could discuss this much further but I'll stop now.
David Lochrin
----------------------
_______________________________________________
Link mailing list
[email protected]
https://mailman.anu.edu.au/mailman/listinfo/link
