A robot may not injure a human being or, through inaction, allow a human being to come to harm
A robot must obey orders given it by human beings except where such orders would conflict with the First Law
A robot must protect its own existence as long as such protection does not conflict with the First or Second Law
Asimov on the Three Laws
The Three Laws are the only way in which rational human beings can deal with robots - or anything else. But when I say that, I always remember (sadly) that human beings are not always rational!
About 300,000 years ago: Homo Sapiens
Homo-sapiens
About 50,000 years ago: Behavioural Modernity
Lascaux cave painting
Human intelligence
Seeing
Uses about 30% of our brain
Optic nerves contain 1 million fibres
Seeing
Being blind
Having low vision
Being colour blind
Hearing
Uses about 3% of our brain
Auditory nerves contain 30,000 fibres
Hearing
Being Deaf
Being hard of hearing
Speaking
Uses about 50% of our brain
Speaking
Being unable to speak
Being unable to communicate using speech
Thinking
Uses most of our brain
Thinking
Being unable to process information
Being unable to remember easily
Being unable to read or understand text
1950: Computing Machinery and Intelligence
Alan Turing
1956: Dartmouth workshop on AI
every aspect of learning or any other feature of intelligence can be so precisely described that a machine can be made to simulate it
Seeing
The process of analysing and understanding images
Hearing
The process of identifying speech and translating it into text
Speaking
The process of translating text into synthetic speech
Thinking
The process of identifying, analysing, understanding, and responding to data
1. A robot may not injure a human being or, through inaction, allow a human being to come to harm
1976: Simulating sight in real-time
Estimated 1,000 MIPs
Cray 1 (fastest computer) capable of 80 - 130 MIPs
2011: Computer vision applications
Averaged between 10,000 and 1 million MIPs
1980s: Moravec's paradox
High-level reasoning requires little computation, but low-level sensory processing requires enormous computational resources
A chair
And another
And one more
2006: Cloud computing
Eric Schmidt
2016: Facebook image recognition
One person in a room close up
2017: Microsoft SeeingAI
2018: Apple's face ID
2. A robot must obey orders given it by human beings except where such orders would conflict with the First Law
1993: Apple voice recognition
1997: Dragon speech recognition
2016: Google's Deep Mind translation
2013: Kinetic Sign Language translation
1769: Speaking machine
Wolfgang von Kempelen's speaking machine
1931: Voder speaking machine
1961: Daisy Bell
1984: Apple MacInTalk
2014: Amazon Echo
3. A robot must protect its own existence as long as such protection does not conflict with the First or Second Law
2013: IBM Watson learns to swear
IBM WatsonUrban Dictionary
2016: Microsoft's Tay.AI learns bad behaviour
MicrosoftTay.AIRacism
2017: Google Translate has gender bias
GoogleGender Bias
2018: Amazon Rekognition mis-identifies US Congress
