When the Turing Test is not enough Functionalist determinations of consciousness and the advent of an authentic machine ethics George Dvorsky Institute for Ethics and Emerging Technologies
Oct 21, 2014
When the Turing Test is not enough
Functionalist determinations of consciousness and the advent of an authentic machine ethics
George DvorskyInstitute for Ethics and Emerging Technologies
Overview
1. Introduction
2. The problem
3. Ethical Implications
4. Solutions
5. Next steps for the citizen scientist
Introduction• Machine consciousness is a
neglected area
• Machine ethics is even further behind
• We need to think about this preemptively
• A complex issue that combines science, philosophy, ethics and law
• A separate issue from robot ethics
What we’re not talking about
“Big Dog”
NOMORAL WORTH
What we’re talking about
Subjective awareness → moral worth
The Problem
• Persistence of vitalism
• Scientific ignorance, defeatism and denial
• Fixation on AI
• Human exceptionalism and substrate chauvinism
• Empiricism versus true scientific understanding
• Persistence of vitalism
• Scientific ignorance, defeatism and denial
• Fixation on AI
• Human exceptionalism and substrate chauvinism
• Empiricism versus true scientific understanding
The Problem
The Problem
• The Turing Test
• A purely behavioral approach
• Conflates intelligence with consciousness
• Inadequately assesses intelligence
• Subject to the anthropomorphic fallacy
• Fails to account for the difficulty in articulating conscious awareness
The ProblemJust because it looks like a duck and
quacks like a duck doesn’t mean it is a duck. Moreover, it doesn’t mean you understand the duck.
“What I cannot create I cannot understand.”- Richard Feynman
This is why you need to build the duck.
Ethical Implications
• AI/AC experimentation
• Human augmentation
• Whole brain emulation/uploads
• Maintaining social cohesion and justice
Solutions
1. Adopt cognitive functionalism as a methodological approach
2. Identify and develop functions sufficient for bringing about subjective awareness in AI
3. Expand protections in the legal realm
Solutions
• Cognitive functionalism as a methodological approach: The proof is in the pudding
• Map the ‘organs of consciousness’ and awareness-specific cognitive function
• Build an AC
• Identify functional analogues in AC
Solutions
• Definition and context setting• Adaptation and learning• Editing• Flagging and debugging• Recruiting and control• Decision-making (executive function)• Analogy forming-function• Metacognitive and self-monitoring
function• Autoprogramming and self-
maintenance function• Definitional and context-setting
function
• Mapping the ‘organs of conscious function’• Bernard Baars (1988)
Solutions
• Brain as state machine• Inner neuron partitioning• Conscious and unconscious states• Perceptual learning and memory• Prediction• Self-Awareness• Representation and meaning• Learning utterances• Learning language• Will• Instinct• Emotion
• Mapping the ‘organs of conscious function’• Igor Aleksander (1995)
• Minimum intelligence• Self-awareness and self-control• A sense of time, future and past• The capability of relating to others• Concern for others• Communication• Control of existence• Curiosity• Change and changeability• Balance of rationality and feeling• Idiosyncrasy• Neocortical functioning
Solutions• Mapping personhood-specific cognitive function
• Joseph Fletcher (1988)
• Identification of functional analogues in AC:• Knowing which mechanisms are
responsible for the emergence of sentience
• Identifying functions responsible for the emergence of self-awareness
• Recognizing these functions in a synthesized context
Solutions
• Expanded protections in the legal realm• Laws to protect machine minds
• Basic fundamental rights as accorded to any person
• Additionally, qualifying AC’s have right to:• Not be shut down• Not be experimented upon• Not have its source code manipulated
against its will• Own its own source code• Have full and unhindered access to its
own source code• Privacy (right to its own internal
mental states)• Self-determination
Solutions
Next steps for thecitizen scientist
• Support the neurosciences• Recognize and promote the concept
of non-human animal sentience and personhood, including the idea that animals are not property
• Advocate for legally binding rights that protect non-human animals
• Oppose the patenting of life, genomes and functional equivalents
• Be prepared to use these legal precedents for when AC emerges
For more info:Web: Institute for Ethics and Emerging Technologies
Blog: Sentient DevelopmentsEmail: [email protected]
Twitter: @GeorgeDvorsky
George DvorskyInstitute for Ethics and Emerging Technologies