HAL 2017 – Interactions between AI and People CWIN | AIE | SF December 7, 2017 Rob Vellinga
Jan 21, 2018
HAL 2017 –Interactions between AI and People
CWIN | AIE | SFDecember 7, 2017Rob Vellinga
2© Capgemini 2017. All rights reserved |Hal 2017 – Interactions between AI and People | Rob Vellinga | 12/7/2017
We’re 80 years into the talk aboutand the reality of Artificial Intelligence
Programing to perform Intellectual Tasks
AI Maturity and Popularity
Advances in AI and Machine Learning
Breaking through with AI & Machine Learning
• 1936 – Alan Turing invented the Universal Turing Machine
• 1949 - Claude Shannon wrote ‘Programming a computer for Playing Chess’
• 1950 – Alan Turing invented the Turing Test, to test intelligence in a computer, requiring that a human being should be unable to distinguish the machine from another human being by using the replies to questions put to both.
• 1956 – Allen Newell and Herbert Simon invented the first ‘thinking machine’-Logic Theorist, a punch card calculator, to produce simulation radar maps.
• 1964 - Danny BoBrow (MIT) shows that computers could Process Natural Language
• 1965 – ELIZA, a program that can hold any discourse in English, simulating the conversation of a psychotherapist
• 1968 – A Space Odyssey byStanley Kubrick
• 1980 – First conference of the American Association of Artificial Intelligence is held
• 1986 - First driverless car is designed by Mercedes Benz
• 1997 – NASA successfully completes the Pathfinder Mission with the first autonomous robotic rover to land on Mars
• 1997 – The first official Robocupsoccer match took place. 40 teams of real and ‘simulated’players
• 1997 – IBM’s Deep Blue wins a chess championship against Gary Kasparov, world chess champion
• 2006 - ‘Machine Reading’ (the autonomous understanding of text)
• 2007 – First database of annotated images to aid visual object recognition and facial recognition
• 2009 - Google starts developing Driverless car
• 2014– Amazon announces Alexa Apple’s Siri switches to neural network Machine Learning
• 2017 - Apple builds Core ML a machine learning framework
Scientific Community accepts concept of AI
3© Capgemini 2017. All rights reserved |Hal 2017 – Interactions between AI and People | Rob Vellinga | 12/7/2017
Why me, why now, why here
People• Anthropometry• Biomechanics• Psychology• Motor Memory• Cognitive Processing• Memory Loading• Patterns• Usability• User-Centered Design
Process• Who are we designing for• What is their mental model• How do they expect the process to
go• How is the process done
manually, supported, augmented, automated
• Define, Design, Develop, Test, Launch
Environments• Places where we spend a lot of time• Places that we spend a lot of effort• Places where we spend a lot of money• Places that are critical to our success• Places that are critical to our safety• Places where the lines between
natural intelligence and artificial intelligence are blurring
Rob Vellinga – Observer, Adopter, Simplify-er
4© Capgemini 2017. All rights reserved |Hal 2017 – Interactions between AI and People | Rob Vellinga | 12/7/2017
Design Thinking starts with empathy, curiosity and watching the world around us.
“Once a Day, Deeply Observe the Ordinary.Good design thinkers observe. Great design thinkers
observe the ordinary.”
Tim Brown - Ideo
5© Capgemini 2017. All rights reserved |Hal 2017 – Interactions between AI and People | Rob Vellinga | 12/7/2017
The Reality of my day…..today.Automation, Analytics, User Experience
6© Capgemini 2017. All rights reserved |Hal 2017 – Interactions between AI and People | Rob Vellinga | 12/7/2017
We’ve been waiting for such a time as this….
AI is not necessarily smarter than a human, but is able to think differently and process LARGE amounts of data.- Kevin Kelley (Wired)
LOW Cost Chips
Cheap computer chips, particularly developed for gaming has resulted in Graphics Processing Units (GPUs)Lowered the cost to explore and enable AI
LIMITLESS Training Sets
Training sets are now able to be created or enable in large, large numbers, thanks to the neural networks.Reduced the time to train and refine, and Increased the scale of deployment of these data sets..
LARGE Scale Neural Networks
Neural networks are now working in large scale.Increased the ability for Deep Learning
1
2
3
7© Capgemini 2017. All rights reserved |Hal 2017 – Interactions between AI and People | Rob Vellinga | 12/7/2017
Five Skills of AI(AI) is similar to our perception of
human intelligence – this isn’t built ona single element, it’s a combination ofsenses, experiences, and knowledge.
– Christopher Stancombe
8© Capgemini 2017. All rights reserved |Hal 2017 – Interactions between AI and People | Rob Vellinga | 12/7/2017
Designing for the Now andthe Not Yet
Autonomous vehicles
Computers recognizing images within photographs
Conversational bots
Recommendation engines
NowDrones for personal or corporate security
Predictive maintenance
Diagnose diseases
Personalized advertising
Not yet
9© Capgemini 2017. All rights reserved |Hal 2017 – Interactions between AI and People | Rob Vellinga | 12/7/2017
The Narrative is that AI is ….
There is no such thing as “an AI.” Enterprises don’t need an “AI.” They need business results in which AI technologies may play a role.
AI is a collection of technologies that can be used in applications, systems and solutions to add specific functional capabilities.
… a product or a thing
“AI” become a generalized marketing term, often with little substance or value.
Very useful, specific functions have been created (speech and image recognition, game playing, fraud and failure prediction), but no general intelligence is in sight, or is singularity (think Hal 9000 or higher).
AI is currently as real as we design it to be.
… pervasive
People define the goals; Technologies execute them.
Technologies (whether AI or not) do not have their own goals that they seek to achieve. Machines execute the programs they have been fed, whether the programs consist of code, data or both. In AI, goal-seeking is an illusion programmed in by people.
Governance and operating parameters define the outcomes that are desired and the guardrails within which the ‘game is played.’
… uncontrollable
AI developers use advanced analytics, special algorithms and large bodies of data to deceive people into believing that their product learns on its own and understands, thinks and empathizes with the user. The tendency is to anthropomorphize technologies. Products, systems, and services with AI are currently as real as we design them to be.
… a simulation and (and replacement) of humanity
10© Capgemini 2017. All rights reserved |Hal 2017 – Interactions between AI and People | Rob Vellinga | 12/7/2017
Design Spectrum
11© Capgemini 2017. All rights reserved |Hal 2017 – Interactions between AI and People | Rob Vellinga | 12/7/2017
UnderstandPersonas, Outcomes, Touchpoints, MoT, Context
ObserveEmpathize, Observe, Ask Questions
Define a POVFrame the problem, Define criteria to evaluate,Be more specific than general
IdeateAsk the right questions,What is the ‘wicked’ question,Need volume and variety of ideas
PrototypeUser Scenarios to tell the story,High fidelity to drive real conversations,Define and Design with representative data.
Test Prototype or POCUser Assessments,Usability Testing, Stakeholder Reviews
User-Centered and Outcome-Based
12© Capgemini 2017. All rights reserved |Hal 2017 – Interactions between AI and People | Rob Vellinga | 12/7/2017 12© Capgemini 2017. All rights reserved |
Evolution of AI Design
2D Internet to 3D InternetAugmented reality moves the user experience from a flat screen to an immersive UX in our 3D world.
Outcome-Based DesignsPlay music, Win the game, Set the temperature, Adjust the room, Know my preferences, Optimize the peformance, Self-maintenance
Facial Recognition and UXNow the machine can react to you and adapt rather than you adapting to the machine.
Recent and Ongoing developments that deepen the relationship and trust between AI and People
Michael E. Porter & James E. Heppelmann
13© Capgemini 2017. All rights reserved |Hal 2017 – Interactions between AI and People | Rob Vellinga | 12/7/2017
Reads facial expressions
Detects emotion
Responds to voice, touch, gestures
Listens....sometimes
Very good at specific tasks
Constantly learning
My AI
A global leader in consulting, technology services and digital transformation,Capgemini is at the forefront of innovation to address the entire breadth of clients’opportunities in the evolving world of cloud, digital and platforms. Building on itsstrong 50-year heritage and deep industry-specific expertise, Capgemini enablesorganizations to realize their business ambitions through an array of services fromstrategy to operations. Capgemini is driven by the conviction that the businessvalue of technology comes from and through people. It is a multicultural companyof 200,000 team members in over 40 countries. The Group reported 2016 globalrevenues of EUR 12.5 billion.
About Capgemini
Learn more about us at
www.capgemini.com
This message contains information that may be privileged or confidential and is the property of the Capgemini Group.
Copyright © 2017 Capgemini. All rights reserved.
People matter, results count.
15© Capgemini 2017. All rights reserved |Hal 2017 – Interactions between AI and People | Rob Vellinga | 12/7/2017
Key TakeawaysHal9000 vs. Hal of 2017
Natural Power to Artificial PowerNatural Intelligence to Artificial Intelligence
Outcomes drive AI Design
2D Internet into 3D Internet