Artificial
intelligence in relation to considerations from the perspective of history of
science, philosophy (of nature) and ethics
Seminar in summer term 2020
Saarland University Informatics Campus
Lecturer:
Dr. Christian Müller, Principal Researcher and Research Fellow at the
German Research Center for Artificial Intelligence, DFKI.
christian.mueller@dfki.de
Assisted and supported by: Iris Merget, Samantha Hubert
The main language of this seminar
will be English. We offer to explain/rephrase some of the content in German,
but this is of course only possible for a small fraction. We will also offer German
sessions for those who understand English but are more comfortable with German.
For those of you who don’t understand English, we cannot recommend the seminar.
In comparison to the discussion
"around AI and ethics", which is currently being conducted by
stakeholders from industry and society in various committees, this seminar is
distinguished by the fact that it aims to get closer to the bottom. We will
approach things from a philosophical perspective and, as the title already
suggests, we will incorporate aspects of the history of science, philosophy of
nature and (this is self-evident) ethics. We will apply these principles in a
more pragmatic part to the discussion "on AI and ethics".
But let's stick to the basics. Here
we will begin by taking a look at the philosophy of evolution. Coming from
there we will derive two views on the nature of humanity: First, a
materialistic-reductionist view, which, in a very condensed form, states that
consciousness is a more or less causal epiphenomenon that, at a certain point
in time, emerged on our highly complex neocortex; and second, a view that sees the
consciousness as the preceding, fundamental principle.
We will look at arguments for both
views.
The following list of topics results
from these preliminary considerations.
·
Introductory topics
on AI and philosophical foundations (short)
Initial definition from our own work. Identification of the basic
characteristics of the AI "components" in relation to relevant questions/problems.
·
Is there knowledge
and if so, what is knowledge?
Epistemology from Plato until today (in Asia and in the "Occident", using
selected examples)
The epistemology conveys essential philosophical foundations
for the seminar and helps us later in the AI definition.
·
What is ethics? Why
is ethics a philosophical topic?
We are doing this to make sure that we all understand
how to distinguish philosophical thinking from practical ethics of the standardisation bodies. In the basics, we will above all consult
the humanist Julian Nida-Rümelin.
·
Basics from the
philosophy of evolution
We will deeper the discussion about materialism versus
alternative perspectives onto the world. We can only speak meaningfully about
AI if we can make a reference to a view on ourselves.
·
What is
intelligence? What about it can be artificial at all?
Part of this complex of topics will
be the two views on the nature of man, which are mentioned above. Here, too, we
will offer a selective historical outline that will provide us with a tool to
sharpen concepts such as perception, understanding, reason, etc. also for their
artificial correspondence (if applicable).
·
The modern dilemma:
the difficult legacy of Newton and Decartes. An
integral view
Behind this lies the question: where does the (deep)
ethical debate come from and what are the problematic positions? Not only Newton and Decartes
play a role here, but also, for example, John Locke on the side of rationalism
and empiricism, and on the other side, for example, German idealism (Kant,
Fichte, Schelling, Hegel). The complex
includes exciting philosophical and scientific battles, such as Goethe's
outburst in connection with the criticism of his theory of colours,
but also forward-looking integral solutions.
·
What does the AI do
particularly in terms of ethical issues (compared to IT in general)?
This is of course an extremely exciting topic, because AI plays an interesting double
(even triple) role here, which we will differentiate and discuss in the seminar.
·
Science-historical
parallels between physics and AI
As is well known, Elon Musk recently
compared the dangers of AI with those of nuclear weapons. Polemical, of course,
but what makes the man think so? We'll explain and put it into perspective.
·
Good AI, bad AI -
does that exist? What distinguishes them both?
Some will say flatland after all. As mentioned at the
beginning, we are here to connect to the current discussion "on AI and
ethics". This is then the point where we will - in short - set the
framework for the results of the seminar with reference to the title.
·
Strong AI: can it
never exist, should it never be allowed to exist, or should it?
Those who have read carefully up to here may make assumptions about which
arguments are being held against each other here.
Due to the Corona crisis, we will
offer the seminar online. Presentations will be recorded and distributed among
the participants. Actual online sessions will we used for discussion, not for
presentation.
The contributions / study
achievements of the seminar will consist of the recorded presentation and a short written report.
Selection of topics will be done on
the basis of the 3 introductionary parts by C.
Müller.
Date |
Discussion of… |
By |
|
Introduction Part I |
C. Müller |
|
Introduction Part II |
C. Müller |
|
Introduction Part III |
C. Müller |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Nagel, T. (2012). Mind and cosmos: why the materialist
neo-Darwinian conception of nature is almost certainly false. Oxford University
Press.
Nagel, Thomas. "What is it like to be a bat?." The philosophical review 83.4 (1974): 435-450.
Wilber, K. (2001). A brief history of everything.
Shambhala Publications.
Wilber, K. (2017). The Religion of Tomorrow: A Vision for
the Future of the Great Traditions-More Inclusive, More Comprehensive, More
Complete. Shambhala Publications. Chicago
Nida-Rümelin, J., & Weidenfeld, N. (2018). Digitaler Humanismus:
eine Ethik für das Zeitalter der künstlichen Intelligenz. Piper ebooks.
Nida-Rümelin, J. (2016). Humanistische Reflexionen. Suhrkamp Verlag.
Bodo Hemprich. Goethes Farbenlehre.
Grundgedanken – Würdigung der Methode. In: Die Drei: Zeitschrift für
Wissenschaft, Kunst und soziales Leben. Ausgabe 11,
November 1982.
Wittgensteinian Philosophy and Advaita
Vedanta: A Survey of the Parallels (2007).
(available as hard copy)
Other authors and sources
Carol Gilligan (ethics of care)
Abraham Maslov
The Upanishades
Great Traditions
Neo Platonism
John Eccles
B.F. Skinner
Locke
Thomas Kuhn
Rupert Sheldrake
Ervin Lazlo
…
1)
How much reading is expected per participant?
Answer: There
are 50+ introductionary slides. From there, I would
estimate another 100-200 textbook pages per participant
2)
How many topics is supposed to be covered by one
participant?
Answer: 1
3)
How many research papers / textbook chapters / articles
does each topic entail and also the intro?
Answer: The
seminar will focus on textbook chapters rather than research articles. The
number of chapters varies depending on the topic. On average, we expect entailment
of 2-3 textbook chapters. See also question 1.
4)
How much work is to be presented and summarized?
We expect a
presentation of about 20 minutes recorded (e.g. using zoom) and distributed
among the participants and lecturers. Then another 20 – 30 minutes discussion
with be held online (synchronously). Finally, we expect a short 2-3 page summary.