Post on 25-Jul-2020
ARTIFICIAL INTELLIGENCE
THE INSIGHTS YOU NEED FROM HARVARD
BUSINESS REVIEW
BY HARVARD BUSINESS REVIEW,
THOMAS H. DAVENPORT, ERIK
BRYNJOLFSSON, ANDREW MCAFEE,
AND H. JAMES WILSON
Contents
Figure 1-1 3
Table 1-1 4
Figure 2-1 5
Figure 2-2 6
Figure 2-3 7
The Value of Collaboration 8
Table 9-1 10
Vision error rate30%
25
20
15
10
Humans
Algorithms
5
02010 2011 2012 2013 2014 2015 2016
F I G U R E 1 - 1
Machines have made real strides in distinguishing among similar- looking categories of images
Source: Electronic Frontier Foundation
054-80639_ch01_4P.indd 8 6/22/19 7:33 AM
3
T A B L E 1 - 1
Supervised learning systemsAs two pioneers in the field, Tom Mitchell and Michael I. Jordan, have noted, most of the recent pro gress in machine learning involves mapping from a set of inputs to a set of outputs. Some examples:
Input X Output Y ApplicationVoice recording Transcript Speech recognition
Historical market data Future market data Trading bots
Photo graph Caption Image tagging
Drug chemical properties Treatment efficacy Pharma R&D
Store transaction details Is the transaction fraudulent?
Fraud detection
Recipe ingredients Customer reviews Food recommendations
Purchase histories Future purchase be hav ior Customer retention
Car locations and speed Traffic flow Traffic lights
Faces Names Face recognition
054-80639_ch01_4P.indd 13 6/22/19 7:33 AM
4
H3R&D AML
Prod
uct r
eadi
ness
Time
Product deliveryH2 H1
F I G U R E 2 - 1
Where AI fits in at Facebook
Source: Facebook
054-80639_ch01_4P.indd 36 6/22/19 7:33 AM
5
Translation
Naturallanguage
Speech
Camera
Computervision
Computerphotography
Visual
Audio
Text
F I G U R E 2 - 2
Applied machine learning
Source: Facebook
054-80639_ch01_4P.indd 41 6/22/19 7:33 AM
6
AI/ML expertise required
Ease of use
Ability to build and customize AI
Self-serve AIFor non-technical users, e.g. Lumos
Reusable enginesFor developers outside of AML, e.g. CLUE
ML algorithmsGeneralizable by discipline
Deep learning frameworkCaffe2
AI backboneFBLearner Flow
Self-serve AIFor non-technical users, e.g. Lumos
Reusable enginesFor developers outside of AML, e.g. CLUE
ML algorithmsGeneralizable by discipline
Deep learning frameworkCaffe2
AI backboneFBLearner Flow
AI backboneFBLearner Flow
Self-serve AIFor non-technical users, e.g. Lumos
Reusable enginesFor developers outside of AML, e.g. CLUE
ML algorithmsGeneralizable by discipline
Deep learning frameworkCaffe2
Less
More
Less
More
Less
More
F I G U R E 2 - 3
Layer cake of AI
Source: Facebook
054-80639_ch01_4P.indd 43 6/22/19 7:33 AM
7
The Value of CollaborationCompanies benefit from optimizing collaboration between
humans and artificial intelligence. Five princi ples can help
them do so: Reimagine business pro cesses; embrace ex-
perimentation/employee involvement; actively direct AI
strategy; responsibly collect data; and redesign work to
incorporate AI and cultivate related employee skills. A sur-
vey of 1,075 companies in 12 industries found that the more
of these princi ples companies adopted, the better their AI
initiatives performed in terms of speed, cost savings, rev-
enues, or other operational mea sures (see the figure).
054-80639_ch01_4P.indd 110 6/22/19 7:33 AM
8
7x
6x
5x
4x
3x
2x
1x
0xPerf
orm
ance
impr
ovem
ent
Number of human-machine collaborationprinciples adopted(O indicates the adoption of only basic,noncollaborative AI)
0 1 2 3 4 5
054-80639_ch01_4P.indd 111 6/22/19 7:33 AM
9
T A B L E 9 - 1
Enhancing per for manceAt organ izations in all kinds of industries, humans and AI are collabo-rating to improve five ele ments of business pro cesses.
Ele mentBusiness pro cess
Com pany or organ ization Type of collaboration
Flexibility Auto manufacturing
Mercedes- Benz Assembly robots work safely alongside humans to customize cars in real time.
Product design Autodesk Software suggests new product design concepts as a designer changes par ameters such as materials, cost, and per for mance requirements.
Software development
Gigster AI helps analyze any type of software proj ect, no matter the size or complexity, enabling humans to quickly estimate the work required, or ga nize experts, and adapt workflows in real time.
Speed Fraud detection
HSBC AI screens credit- and debit- card transactions to instantly approve legitimate ones while flagging ques-tionable ones for humans to evaluate.
Cancer treatment
Roche AI aggregates patient data from disparate IT systems, speeding collaboration among specialists.
Public safety Singapore government
Video analytics during public events predicts crowd be-hav ior, helping responders address security incidents rapidly.
Scale Recruiting Unilever Automated applicant screening dramatically expands the pool of quali-fied candidates for hiring man ag ers to evaluate.
054-80639_ch01_4P.indd 124 6/22/19 7:33 AM
10
Ele mentBusiness pro cess
Com pany or organ ization Type of collaboration
Scale (conti nued)
Customer ser vice
Virgin Trains Bot responds to basic cus-tomer requests, doubling the volume handled and freeing humans to address more- complex issues.
Casino management
GGH Morowitz Computer- vision system helps humans continuously monitor every gaming table in a casino.
Decision making
Equipment maintenance
General Electric
“Digital twins” and Predix diagnostic application pro-vide techs with tailored rec-ommendations for machine maintenance.
Financial ser vices
Morgan Stanley Robo- advisers offer clients a range of investment options based on real- time market information.
Disease prediction
Icahn School of Medicine at Mount Sinai
Deep Patient system helps doctors predict patients’ risk of specific disease, allowing preventive intervention.
Personal-ization
Guest experience
Carnival Corporation
Wearable AI device streamlines the logistics of cruise- ship activities and an-ticipates guest preferences, facilitating tailored staff support.
Health care Pfizer Wearable sensors for Par-kinson’s patients track symptoms 24/7, allowing customized treatment.
Retail fashion Stitch Fix AI analyzes customer data to advise human stylists, who give customers individualized clothing and styling recommendations.
054-80639_ch01_4P.indd 125 6/22/19 7:33 AM
11