artificial intelligence - u3asites.org.uk · intelligence 1956: dartford college conference, new...
TRANSCRIPT
Artificial Intelligence
…and the Social Implications
Artificial Intelligence (AI)
What is AI?
History and Development of AI
AI in Everyday Life
Human v Artificial Intelligence
The Future
Wider Implications
What is Artificial Intelligence
Term applied when:
a machine mimics the cognitive functions that humans associate with other human minds, such as learning, reasoning and problem-solving
Machine acts as a rational agent
…anything that:
has clear preferences and makes decisions
models uncertainty (by assessing options and variables)
perceives its environment and takes actions that maximize its chance of success at some goal
Rational Agent
1940-50s: Alan Turing began pioneering work into computer intelligence
1956: Dartford College Conference, New Hampshire, US
John McCarthy coined phrase “Artificial Intelligence” and designed/developed LISP programming language
History and Development of AI
McCarthy’s Proposal
“We propose that a 2 month, 10 man study of artificial
intelligence be carried out during the summer of 1956 at
Dartmouth College in Hanover, New Hampshire. The study
is to proceed on the basis of the conjecture that every aspect
of learning or any other feature of intelligence can in
principle be so precisely described that a machine can be
made to simulate it. An attempt will be made to find how to
make machines use language, form abstractions and
concepts, solve kinds of problems now reserved for humans,
and improve themselves”
1952-62: Arthur Samuel wrote first AI game program (checkers)
1950’s-60’s: Semantic nets developed – represents semantic relations between concepts. This is often used as a form of knowledge representation.
History and Development of AI
Semantic Net – simple example
Lisp Semantic Network
(defun *database* ()'((canary (is-a bird)
(color yellow)(size small))
(penguin (is-a bird)(movement swim))
(bird (is-a vertebrate)(has-part wings)(reproduction egg-laying))))
1965: Eliza, one of the first “Chatterbots”
1996: IBM Deep Blue v. Garry Kasparov at chess
History and Development of AI
AI and the Turing Test
Named after Alan Turing
English mathematician/codebreaker
“A computer would deserve to be called intelligent if it could deceive a human into believing that it was human”
• In Turing's test, if human is unable to consistently determine whether an answer has been given by a computer or by another human being, then the computer is considered to have "passed" the test
• Turing’s paper Computing Machinery and Intelligence (1950) - predicted by Year 2000, machines would be capable of fooling 30% of humans after 5mins of questioning
The Turing Test
1960’s Chatterbots
Also known as talkbot, chatbot, bot, chatterbox, Artificial Conversational Entity.
In essence, a computer program that can conduct a conversation with a human via voice recognition/ simulation or textual methods (screen/keyboard).
Chatterbot programs designed to simulate how a human would behave in conversation.
Chatterbots
Chatterbots can use sophisticated language processing systems, but many simpler systems scan for keywords
For example, the word “Mother”
₋ May formulate a response based on known associations within a database
₋ Mother links to family… Response = “Tell me more about your family”
₋ This seems like a human response
Who is Eugene Goostman?
• Competed in a number of Turing Test contests
• June 2012 - 29% judges thought Goostman was human
• June 2014 - 33% judges thought it was human
• Organisers claimed Eugene had passed Turing's test…..
• Chatterbot, developed in St Petersburg, portrayed as 13-yr-old Ukrainian boy
Validity of Goostman's test result called into question:
• 13 yr old expected to lack general knowledge and display
poor grammar etc
• Goostman made use of personality quirks and humour as an
attempt to misdirect users from its non-human tendencies
Artificial Intelligence
in Everyday Life
Online Assistants
Automated online assistants use integrated chatterbot systems….
These can engage in “small talk” and casual conversations unrelated to their primary purpose.
Many large companies (Lloyds Bank, RBS, Renault, Citroën) use automated online assistants to provide a first point of contact with clients rather than call centres.
Google Now, Siri & Cortana
Intelligent digital personal assistants that operate on various platforms.
process huge volumes of data from a variety of sources
learn about their users
become more effective in helping their users organise and track their information.
AI and Digital Personal Assistants
AI in Google Now, Siri and Cortana
These apps collect information on your requests and use that information to :
Better recognize your speech
Deliver the results that are tailored to your
preferences.
Microsoft says Cortana “continually learns about its user” so that it eventually develops the ability to anticipate users’ needs.
Smart Cars
Self-driving cars are becoming reality eg Google’s self-driving car project and Tesla’s “autopilot” car
Tesla plans new truck, bus and car-sharing systems
Algorithms now being developed by Google to enable self-driving cars to learn to drive like humans - through experience
Google will be testing the intelligence used in advanced video games
So that, eventually, car will be able to “look” at the road ahead and make decisions based on what it sees, helping it learn in the process.
Tesla car isn’t this advanced, but it’s already on the road.
Smart Cars
Smart Home Devices
Smart home devices now able to learn behaviour patterns and routines
Can help save money, time and energy usage….
turning your oven on when you leave work
thermostat knows when you’re home and adjusts the temperature accordingly
lights around your home adjust based on where you are and what you’re doing
Potential of AI in smart homes is huge
AI in Medicine
Numerous examples emerging:
Used in clinical decision support systems for medical diagnosis
Computer-aided interpretation of medical images
AI applied to Electromagnetic Radiation Images (EMR/MRI) and Computed Tomography (CT), and by applying knowledge base of typical images can highlight areas of irregularity, such as possible diseases/tumours
Used in heart sound analysis
Drones for recognisance
Developing autonomous weapons that use AI to locate, track and destroy their targets
Some believe these are more efficient, reliable and less flawed than humans. …don’t need sleep, rest, food, recovery… more easily replaced
Could they be developed to be more ethical?
But evidence from drone strikes paints a different picture
Some nations are seeking ban on use of AI as killing machines
AI and RoboticsAI and the Military
AI and RoboticsAI and robotics can be used for positive purposes ...
Autonomous cars predicted to save 30,000 deaths on US roads per year
Make our roads, factories, mines and ports safer & more efficient
Help blind people to explore new environments
Make our lives healthier, wealthier and happier
In military context, AI/robots can be used to clear minefields, bring supplies in through dangerous routes, and analyse huge amounts of data and intelligence.
…..But should they be used to kill?
Human Intelligence
versus
Artificial Intelligence
Possible Pros Possible Cons
Intuition, common sense,judgement, belief systems
Flawed and fallibleCan be irrational, impetuous
Creativity and in novel situations, can apply unrelated knowledge to solve complex problems
Limited knowledge basesUnable to store large amounts of data in memory
Plausible reasoning and critical thinking
Relatively slow at processing data
Effective human to human communicationsTraits - empathy, tenderness
Anger, frustration, jealousy
Emotional intelligence Emotional hurt, revenge
Humans
Possible Pros Possible Cons
Ability to simulate human behaviour and cognitive processes
No common senseCannot easily handle “mixed”knowledge or novel problems
Can capture and preserve human expertise
High development costs
Can capture, store and process large amounts of data rapidly… Good at frequent, high volume tasks
Ethical and moral concerns especially military applications
Non-emotional, objective, rational Legal issues… responsibility/accountability?
Artificial/Computer intelligence
Artificial Intelligence
….The Future
The Future
AI - a radical step change in our technology, limited only by our own imaginations
Autonomous agents - perceive, decide and act on their own
Raises the possibility that AI agents could take unanticipated actions beyond our control
AI - Wider Implications
Smart AI artifacts are increasingly being deployed
Their widespread deployment will have profound consequences for human society:
moral and ethical
psychological
social
economic
legal……… let’s discuss!
Lisp Example of Praise in Response
(defmacro format-ans-str (op-str op1 op2)
`(format t "~%~2d ~a ~2d = " ,op1 ,op-str ,op2))
(defun kudos (op2)
(let ((kudos-str "")
(kudos))
(setf kudos-str
'("~t Well Done!"
"~t Good Job!"
"~t Great!"
"~t Keep it up!"
"~t You're something, you know"))
(setf kudos (nth (random (length kudos-str)) kudos-str ))
(update-mistake-counter op2 -1)
(format t kudos)))
A computer-generated poem, accepted for publication:
A home transformed by the lightning the balanced alcoves smother this insatiable earth of a planet, Earth. They attacked it with mechanical horns because they love you, love, in fire and wind. You say, what is the time waiting for in its spring? I tell you it is waiting for your branch that flows, because you are a sweet-smelling diamond architecture that does not know why it grows.
Magenta, Google’s art app (in the style of Van Gogh?)