SlideShare a Scribd company logo
Recognizing the cultural bias in Artificial Intelligence
Camille Eddy
ME student, Boise State
Machine Learning Intern, HP Inc.
@NikkyMill
Where does AI exist in our lives?
Addressing Sexism:
Relationships that are common in our language appear
in machine translations
father : doctor :: mother : x
x = nurse
man : computer programmer ::
woman : x
x = homemaker
MIT Technology Review – Word2vec
Addressing Racial Bias:
The way we use algorithms
can have serious
consequences for our
personal knowledge
Addressing Search Bias:
Filter Bubbles
Why is it this way?
Who develops the technology… Who uses the technology…
This prevents us from having a ‘perfect’ algorithm for every user….
How do we fix it?
make it better
/
Accuracy: _______
Explainable AI (XAI)
… And encourage more people to create and consume
We need more organizations
and services to combat bias!
Recognizing the cultural bias in Artificial Intelligence

More Related Content

Viewers also liked

Trabajo en Equipo, un recurso del liderazgo - Ana María Ribera (HUMAN VALUE)
Trabajo en Equipo, un recurso del liderazgo - Ana María Ribera (HUMAN VALUE)Trabajo en Equipo, un recurso del liderazgo - Ana María Ribera (HUMAN VALUE)
Trabajo en Equipo, un recurso del liderazgo - Ana María Ribera (HUMAN VALUE)gdgsantacruz
 
Los Riesgos en Emprendimientos de TI - Claudia Araujo Michel (YANAPTI CORP)
Los Riesgos en Emprendimientos de TI - Claudia Araujo Michel (YANAPTI CORP)Los Riesgos en Emprendimientos de TI - Claudia Araujo Michel (YANAPTI CORP)
Los Riesgos en Emprendimientos de TI - Claudia Araujo Michel (YANAPTI CORP)gdgsantacruz
 
Неделя профилактики употребления психоактивных веществ «Независимое детство»
Неделя профилактики употребления психоактивных веществ «Независимое детство»Неделя профилактики употребления психоактивных веществ «Независимое детство»
Неделя профилактики употребления психоактивных веществ «Независимое детство»yul7059
 
Localizing Your eLearning Course
Localizing Your eLearning CourseLocalizing Your eLearning Course
Localizing Your eLearning CourseCTS LanguageLink
 
Tecnologias educacionais e tecnologias da informação e comunicação
Tecnologias educacionais e tecnologias da informação e comunicaçãoTecnologias educacionais e tecnologias da informação e comunicação
Tecnologias educacionais e tecnologias da informação e comunicaçãoFaculdade Metropolitanas Unidas - FMU
 
UNICO Riviera Maya
UNICO Riviera MayaUNICO Riviera Maya
UNICO Riviera Mayachglat
 
Medicion de la productividad del valor agregado
Medicion de la productividad del valor agregadoMedicion de la productividad del valor agregado
Medicion de la productividad del valor agregadoloreanny vasquez
 
Innovación y desarrollo tecnologico 3°1 tv Sugey Corchado
Innovación y desarrollo tecnologico 3°1 tv Sugey CorchadoInnovación y desarrollo tecnologico 3°1 tv Sugey Corchado
Innovación y desarrollo tecnologico 3°1 tv Sugey CorchadoSugey Corchado
 
Guía práctica 2017.matemática 3.silvia torre
Guía práctica 2017.matemática 3.silvia torreGuía práctica 2017.matemática 3.silvia torre
Guía práctica 2017.matemática 3.silvia torreDon Augusto
 
Presentation about curriculum
Presentation about curriculumPresentation about curriculum
Presentation about curriculumIram Naz
 

Viewers also liked (13)

Trabajo en Equipo, un recurso del liderazgo - Ana María Ribera (HUMAN VALUE)
Trabajo en Equipo, un recurso del liderazgo - Ana María Ribera (HUMAN VALUE)Trabajo en Equipo, un recurso del liderazgo - Ana María Ribera (HUMAN VALUE)
Trabajo en Equipo, un recurso del liderazgo - Ana María Ribera (HUMAN VALUE)
 
Los Riesgos en Emprendimientos de TI - Claudia Araujo Michel (YANAPTI CORP)
Los Riesgos en Emprendimientos de TI - Claudia Araujo Michel (YANAPTI CORP)Los Riesgos en Emprendimientos de TI - Claudia Araujo Michel (YANAPTI CORP)
Los Riesgos en Emprendimientos de TI - Claudia Araujo Michel (YANAPTI CORP)
 
Неделя профилактики употребления психоактивных веществ «Независимое детство»
Неделя профилактики употребления психоактивных веществ «Независимое детство»Неделя профилактики употребления психоактивных веществ «Независимое детство»
Неделя профилактики употребления психоактивных веществ «Независимое детство»
 
Localizing Your eLearning Course
Localizing Your eLearning CourseLocalizing Your eLearning Course
Localizing Your eLearning Course
 
BSW Salem State Project
BSW Salem State ProjectBSW Salem State Project
BSW Salem State Project
 
Tecnologias educacionais e tecnologias da informação e comunicação
Tecnologias educacionais e tecnologias da informação e comunicaçãoTecnologias educacionais e tecnologias da informação e comunicação
Tecnologias educacionais e tecnologias da informação e comunicação
 
UNICO Riviera Maya
UNICO Riviera MayaUNICO Riviera Maya
UNICO Riviera Maya
 
Una llamada al amor tony de mello
Una llamada al amor   tony de melloUna llamada al amor   tony de mello
Una llamada al amor tony de mello
 
Medicion de la productividad del valor agregado
Medicion de la productividad del valor agregadoMedicion de la productividad del valor agregado
Medicion de la productividad del valor agregado
 
Innovación y desarrollo tecnologico 3°1 tv Sugey Corchado
Innovación y desarrollo tecnologico 3°1 tv Sugey CorchadoInnovación y desarrollo tecnologico 3°1 tv Sugey Corchado
Innovación y desarrollo tecnologico 3°1 tv Sugey Corchado
 
Guía práctica 2017.matemática 3.silvia torre
Guía práctica 2017.matemática 3.silvia torreGuía práctica 2017.matemática 3.silvia torre
Guía práctica 2017.matemática 3.silvia torre
 
Sodio caso clínico
Sodio caso clínicoSodio caso clínico
Sodio caso clínico
 
Presentation about curriculum
Presentation about curriculumPresentation about curriculum
Presentation about curriculum
 

Similar to Recognizing the cultural bias in Artificial Intelligence

Research & Business about Artificial Intelligence: A Point of View
Research & Business about Artificial Intelligence: A Point of ViewResearch & Business about Artificial Intelligence: A Point of View
Research & Business about Artificial Intelligence: A Point of ViewPietro Leo
 
Artificial Intelligence AI in Libraries Training for Innovation Webinar
Artificial Intelligence  AI in Libraries Training for Innovation WebinarArtificial Intelligence  AI in Libraries Training for Innovation Webinar
Artificial Intelligence AI in Libraries Training for Innovation WebinarSaid Ali Said
 
Navigating the AI Era Three Pathways to Long-Term Relevance for Every Individ...
Navigating the AI Era Three Pathways to Long-Term Relevance for Every Individ...Navigating the AI Era Three Pathways to Long-Term Relevance for Every Individ...
Navigating the AI Era Three Pathways to Long-Term Relevance for Every Individ...🔷Tisha Jackson🔷
 
The Evolution of Generative Artificial Intelligence What Lies Ahead.pdf
The Evolution of Generative Artificial Intelligence What Lies Ahead.pdfThe Evolution of Generative Artificial Intelligence What Lies Ahead.pdf
The Evolution of Generative Artificial Intelligence What Lies Ahead.pdfTop Trends
 
Revolutionizing L&D: Harnessing the Power of AI to Empower Tomorrow's Workforce
Revolutionizing L&D: Harnessing the Power of AI to Empower Tomorrow's WorkforceRevolutionizing L&D: Harnessing the Power of AI to Empower Tomorrow's Workforce
Revolutionizing L&D: Harnessing the Power of AI to Empower Tomorrow's WorkforceStella Lee
 
The View from Here and Here: Making the Invisible Visible in the Hypertextual...
The View from Here and Here: Making the Invisible Visible in the Hypertextual...The View from Here and Here: Making the Invisible Visible in the Hypertextual...
The View from Here and Here: Making the Invisible Visible in the Hypertextual...Michelle Ferrier
 
artificial intelligence - in need of an ethical layer?
artificial intelligence - in need of an ethical layer?artificial intelligence - in need of an ethical layer?
artificial intelligence - in need of an ethical layer?Inge de Waard
 
Introducing Computational Creativity
Introducing Computational CreativityIntroducing Computational Creativity
Introducing Computational CreativityTony Veale
 
Social Machines Oxford Hendler
Social Machines Oxford HendlerSocial Machines Oxford Hendler
Social Machines Oxford HendlerJames Hendler
 
Hacking Gender Inclusion (Aug17)
Hacking Gender Inclusion (Aug17)Hacking Gender Inclusion (Aug17)
Hacking Gender Inclusion (Aug17)Daniele Fiandaca
 
ARTIFICIAL INTELIGENCE
ARTIFICIAL INTELIGENCEARTIFICIAL INTELIGENCE
ARTIFICIAL INTELIGENCECYMAX
 
Health Essay Questions. Online assignment writing service.
Health Essay Questions. Online assignment writing service.Health Essay Questions. Online assignment writing service.
Health Essay Questions. Online assignment writing service.Umon Kinneberg
 
Why do we need Inclusive AI
Why do we need Inclusive AIWhy do we need Inclusive AI
Why do we need Inclusive AIBharat Krish
 
Writing as academic practice short.pdf
Writing as academic practice short.pdfWriting as academic practice short.pdf
Writing as academic practice short.pdfHelen Beetham
 
Defining the boundary for AI research in Intelligent Systems Dec 2021
Defining the boundary for AI research in Intelligent Systems Dec  2021Defining the boundary for AI research in Intelligent Systems Dec  2021
Defining the boundary for AI research in Intelligent Systems Dec 2021Parasuram Balasubramanian
 
Oas keynote 10 2019
Oas keynote 10 2019Oas keynote 10 2019
Oas keynote 10 2019Jerome Glenn
 
Introduction to ArtificiaI Intelligence in Higher Education
Introduction to ArtificiaI Intelligence in Higher EducationIntroduction to ArtificiaI Intelligence in Higher Education
Introduction to ArtificiaI Intelligence in Higher Educationpboyjonauth
 

Similar to Recognizing the cultural bias in Artificial Intelligence (20)

Research & Business about Artificial Intelligence: A Point of View
Research & Business about Artificial Intelligence: A Point of ViewResearch & Business about Artificial Intelligence: A Point of View
Research & Business about Artificial Intelligence: A Point of View
 
Artificial Intelligence AI in Libraries Training for Innovation Webinar
Artificial Intelligence  AI in Libraries Training for Innovation WebinarArtificial Intelligence  AI in Libraries Training for Innovation Webinar
Artificial Intelligence AI in Libraries Training for Innovation Webinar
 
Artificial Intelligence AI in Libraries Training for Innovation Webinar
Artificial Intelligence  AI in Libraries Training for Innovation WebinarArtificial Intelligence  AI in Libraries Training for Innovation Webinar
Artificial Intelligence AI in Libraries Training for Innovation Webinar
 
Navigating the AI Era Three Pathways to Long-Term Relevance for Every Individ...
Navigating the AI Era Three Pathways to Long-Term Relevance for Every Individ...Navigating the AI Era Three Pathways to Long-Term Relevance for Every Individ...
Navigating the AI Era Three Pathways to Long-Term Relevance for Every Individ...
 
The Evolution of Generative Artificial Intelligence What Lies Ahead.pdf
The Evolution of Generative Artificial Intelligence What Lies Ahead.pdfThe Evolution of Generative Artificial Intelligence What Lies Ahead.pdf
The Evolution of Generative Artificial Intelligence What Lies Ahead.pdf
 
Revolutionizing L&D: Harnessing the Power of AI to Empower Tomorrow's Workforce
Revolutionizing L&D: Harnessing the Power of AI to Empower Tomorrow's WorkforceRevolutionizing L&D: Harnessing the Power of AI to Empower Tomorrow's Workforce
Revolutionizing L&D: Harnessing the Power of AI to Empower Tomorrow's Workforce
 
The View from Here and Here: Making the Invisible Visible in the Hypertextual...
The View from Here and Here: Making the Invisible Visible in the Hypertextual...The View from Here and Here: Making the Invisible Visible in the Hypertextual...
The View from Here and Here: Making the Invisible Visible in the Hypertextual...
 
artificial intelligence - in need of an ethical layer?
artificial intelligence - in need of an ethical layer?artificial intelligence - in need of an ethical layer?
artificial intelligence - in need of an ethical layer?
 
Introducing Computational Creativity
Introducing Computational CreativityIntroducing Computational Creativity
Introducing Computational Creativity
 
Social Machines Oxford Hendler
Social Machines Oxford HendlerSocial Machines Oxford Hendler
Social Machines Oxford Hendler
 
Ppt4
Ppt4Ppt4
Ppt4
 
Hacking Gender Inclusion (Aug17)
Hacking Gender Inclusion (Aug17)Hacking Gender Inclusion (Aug17)
Hacking Gender Inclusion (Aug17)
 
ARTIFICIAL INTELIGENCE
ARTIFICIAL INTELIGENCEARTIFICIAL INTELIGENCE
ARTIFICIAL INTELIGENCE
 
Health Essay Questions. Online assignment writing service.
Health Essay Questions. Online assignment writing service.Health Essay Questions. Online assignment writing service.
Health Essay Questions. Online assignment writing service.
 
Why do we need Inclusive AI
Why do we need Inclusive AIWhy do we need Inclusive AI
Why do we need Inclusive AI
 
Ethics in Technology Handout
Ethics in Technology HandoutEthics in Technology Handout
Ethics in Technology Handout
 
Writing as academic practice short.pdf
Writing as academic practice short.pdfWriting as academic practice short.pdf
Writing as academic practice short.pdf
 
Defining the boundary for AI research in Intelligent Systems Dec 2021
Defining the boundary for AI research in Intelligent Systems Dec  2021Defining the boundary for AI research in Intelligent Systems Dec  2021
Defining the boundary for AI research in Intelligent Systems Dec 2021
 
Oas keynote 10 2019
Oas keynote 10 2019Oas keynote 10 2019
Oas keynote 10 2019
 
Introduction to ArtificiaI Intelligence in Higher Education
Introduction to ArtificiaI Intelligence in Higher EducationIntroduction to ArtificiaI Intelligence in Higher Education
Introduction to ArtificiaI Intelligence in Higher Education
 

Recently uploaded

Future Visions: Predictions to Guide and Time Tech Innovation, Peter Udo Diehl
Future Visions: Predictions to Guide and Time Tech Innovation, Peter Udo DiehlFuture Visions: Predictions to Guide and Time Tech Innovation, Peter Udo Diehl
Future Visions: Predictions to Guide and Time Tech Innovation, Peter Udo DiehlPeter Udo Diehl
 
Search and Society: Reimagining Information Access for Radical Futures
Search and Society: Reimagining Information Access for Radical FuturesSearch and Society: Reimagining Information Access for Radical Futures
Search and Society: Reimagining Information Access for Radical FuturesBhaskar Mitra
 
Essentials of Automations: Optimizing FME Workflows with Parameters
Essentials of Automations: Optimizing FME Workflows with ParametersEssentials of Automations: Optimizing FME Workflows with Parameters
Essentials of Automations: Optimizing FME Workflows with ParametersSafe Software
 
FIDO Alliance Osaka Seminar: FIDO Security Aspects.pdf
FIDO Alliance Osaka Seminar: FIDO Security Aspects.pdfFIDO Alliance Osaka Seminar: FIDO Security Aspects.pdf
FIDO Alliance Osaka Seminar: FIDO Security Aspects.pdfFIDO Alliance
 
Designing Great Products: The Power of Design and Leadership by Chief Designe...
Designing Great Products: The Power of Design and Leadership by Chief Designe...Designing Great Products: The Power of Design and Leadership by Chief Designe...
Designing Great Products: The Power of Design and Leadership by Chief Designe...Product School
 
Assuring Contact Center Experiences for Your Customers With ThousandEyes
Assuring Contact Center Experiences for Your Customers With ThousandEyesAssuring Contact Center Experiences for Your Customers With ThousandEyes
Assuring Contact Center Experiences for Your Customers With ThousandEyesThousandEyes
 
Dev Dives: Train smarter, not harder – active learning and UiPath LLMs for do...
Dev Dives: Train smarter, not harder – active learning and UiPath LLMs for do...Dev Dives: Train smarter, not harder – active learning and UiPath LLMs for do...
Dev Dives: Train smarter, not harder – active learning and UiPath LLMs for do...UiPathCommunity
 
Quantum Computing: Current Landscape and the Future Role of APIs
Quantum Computing: Current Landscape and the Future Role of APIsQuantum Computing: Current Landscape and the Future Role of APIs
Quantum Computing: Current Landscape and the Future Role of APIsVlad Stirbu
 
ODC, Data Fabric and Architecture User Group
ODC, Data Fabric and Architecture User GroupODC, Data Fabric and Architecture User Group
ODC, Data Fabric and Architecture User GroupCatarinaPereira64715
 
"Impact of front-end architecture on development cost", Viktor Turskyi
"Impact of front-end architecture on development cost", Viktor Turskyi"Impact of front-end architecture on development cost", Viktor Turskyi
"Impact of front-end architecture on development cost", Viktor TurskyiFwdays
 
Mission to Decommission: Importance of Decommissioning Products to Increase E...
Mission to Decommission: Importance of Decommissioning Products to Increase E...Mission to Decommission: Importance of Decommissioning Products to Increase E...
Mission to Decommission: Importance of Decommissioning Products to Increase E...Product School
 
IOS-PENTESTING-BEGINNERS-PRACTICAL-GUIDE-.pptx
IOS-PENTESTING-BEGINNERS-PRACTICAL-GUIDE-.pptxIOS-PENTESTING-BEGINNERS-PRACTICAL-GUIDE-.pptx
IOS-PENTESTING-BEGINNERS-PRACTICAL-GUIDE-.pptxAbida Shariff
 
Neuro-symbolic is not enough, we need neuro-*semantic*
Neuro-symbolic is not enough, we need neuro-*semantic*Neuro-symbolic is not enough, we need neuro-*semantic*
Neuro-symbolic is not enough, we need neuro-*semantic*Frank van Harmelen
 
To Graph or Not to Graph Knowledge Graph Architectures and LLMs
To Graph or Not to Graph Knowledge Graph Architectures and LLMsTo Graph or Not to Graph Knowledge Graph Architectures and LLMs
To Graph or Not to Graph Knowledge Graph Architectures and LLMsPaul Groth
 
FIDO Alliance Osaka Seminar: Overview.pdf
FIDO Alliance Osaka Seminar: Overview.pdfFIDO Alliance Osaka Seminar: Overview.pdf
FIDO Alliance Osaka Seminar: Overview.pdfFIDO Alliance
 
Elevating Tactical DDD Patterns Through Object Calisthenics
Elevating Tactical DDD Patterns Through Object CalisthenicsElevating Tactical DDD Patterns Through Object Calisthenics
Elevating Tactical DDD Patterns Through Object CalisthenicsDorra BARTAGUIZ
 
Knowledge engineering: from people to machines and back
Knowledge engineering: from people to machines and backKnowledge engineering: from people to machines and back
Knowledge engineering: from people to machines and backElena Simperl
 
State of ICS and IoT Cyber Threat Landscape Report 2024 preview
State of ICS and IoT Cyber Threat Landscape Report 2024 previewState of ICS and IoT Cyber Threat Landscape Report 2024 preview
State of ICS and IoT Cyber Threat Landscape Report 2024 previewPrayukth K V
 
Accelerate your Kubernetes clusters with Varnish Caching
Accelerate your Kubernetes clusters with Varnish CachingAccelerate your Kubernetes clusters with Varnish Caching
Accelerate your Kubernetes clusters with Varnish CachingThijs Feryn
 

Recently uploaded (20)

Future Visions: Predictions to Guide and Time Tech Innovation, Peter Udo Diehl
Future Visions: Predictions to Guide and Time Tech Innovation, Peter Udo DiehlFuture Visions: Predictions to Guide and Time Tech Innovation, Peter Udo Diehl
Future Visions: Predictions to Guide and Time Tech Innovation, Peter Udo Diehl
 
Search and Society: Reimagining Information Access for Radical Futures
Search and Society: Reimagining Information Access for Radical FuturesSearch and Society: Reimagining Information Access for Radical Futures
Search and Society: Reimagining Information Access for Radical Futures
 
Essentials of Automations: Optimizing FME Workflows with Parameters
Essentials of Automations: Optimizing FME Workflows with ParametersEssentials of Automations: Optimizing FME Workflows with Parameters
Essentials of Automations: Optimizing FME Workflows with Parameters
 
FIDO Alliance Osaka Seminar: FIDO Security Aspects.pdf
FIDO Alliance Osaka Seminar: FIDO Security Aspects.pdfFIDO Alliance Osaka Seminar: FIDO Security Aspects.pdf
FIDO Alliance Osaka Seminar: FIDO Security Aspects.pdf
 
Designing Great Products: The Power of Design and Leadership by Chief Designe...
Designing Great Products: The Power of Design and Leadership by Chief Designe...Designing Great Products: The Power of Design and Leadership by Chief Designe...
Designing Great Products: The Power of Design and Leadership by Chief Designe...
 
Assuring Contact Center Experiences for Your Customers With ThousandEyes
Assuring Contact Center Experiences for Your Customers With ThousandEyesAssuring Contact Center Experiences for Your Customers With ThousandEyes
Assuring Contact Center Experiences for Your Customers With ThousandEyes
 
Dev Dives: Train smarter, not harder – active learning and UiPath LLMs for do...
Dev Dives: Train smarter, not harder – active learning and UiPath LLMs for do...Dev Dives: Train smarter, not harder – active learning and UiPath LLMs for do...
Dev Dives: Train smarter, not harder – active learning and UiPath LLMs for do...
 
Quantum Computing: Current Landscape and the Future Role of APIs
Quantum Computing: Current Landscape and the Future Role of APIsQuantum Computing: Current Landscape and the Future Role of APIs
Quantum Computing: Current Landscape and the Future Role of APIs
 
ODC, Data Fabric and Architecture User Group
ODC, Data Fabric and Architecture User GroupODC, Data Fabric and Architecture User Group
ODC, Data Fabric and Architecture User Group
 
"Impact of front-end architecture on development cost", Viktor Turskyi
"Impact of front-end architecture on development cost", Viktor Turskyi"Impact of front-end architecture on development cost", Viktor Turskyi
"Impact of front-end architecture on development cost", Viktor Turskyi
 
Ransomware Mallox [EN].pdf
Ransomware         Mallox       [EN].pdfRansomware         Mallox       [EN].pdf
Ransomware Mallox [EN].pdf
 
Mission to Decommission: Importance of Decommissioning Products to Increase E...
Mission to Decommission: Importance of Decommissioning Products to Increase E...Mission to Decommission: Importance of Decommissioning Products to Increase E...
Mission to Decommission: Importance of Decommissioning Products to Increase E...
 
IOS-PENTESTING-BEGINNERS-PRACTICAL-GUIDE-.pptx
IOS-PENTESTING-BEGINNERS-PRACTICAL-GUIDE-.pptxIOS-PENTESTING-BEGINNERS-PRACTICAL-GUIDE-.pptx
IOS-PENTESTING-BEGINNERS-PRACTICAL-GUIDE-.pptx
 
Neuro-symbolic is not enough, we need neuro-*semantic*
Neuro-symbolic is not enough, we need neuro-*semantic*Neuro-symbolic is not enough, we need neuro-*semantic*
Neuro-symbolic is not enough, we need neuro-*semantic*
 
To Graph or Not to Graph Knowledge Graph Architectures and LLMs
To Graph or Not to Graph Knowledge Graph Architectures and LLMsTo Graph or Not to Graph Knowledge Graph Architectures and LLMs
To Graph or Not to Graph Knowledge Graph Architectures and LLMs
 
FIDO Alliance Osaka Seminar: Overview.pdf
FIDO Alliance Osaka Seminar: Overview.pdfFIDO Alliance Osaka Seminar: Overview.pdf
FIDO Alliance Osaka Seminar: Overview.pdf
 
Elevating Tactical DDD Patterns Through Object Calisthenics
Elevating Tactical DDD Patterns Through Object CalisthenicsElevating Tactical DDD Patterns Through Object Calisthenics
Elevating Tactical DDD Patterns Through Object Calisthenics
 
Knowledge engineering: from people to machines and back
Knowledge engineering: from people to machines and backKnowledge engineering: from people to machines and back
Knowledge engineering: from people to machines and back
 
State of ICS and IoT Cyber Threat Landscape Report 2024 preview
State of ICS and IoT Cyber Threat Landscape Report 2024 previewState of ICS and IoT Cyber Threat Landscape Report 2024 preview
State of ICS and IoT Cyber Threat Landscape Report 2024 preview
 
Accelerate your Kubernetes clusters with Varnish Caching
Accelerate your Kubernetes clusters with Varnish CachingAccelerate your Kubernetes clusters with Varnish Caching
Accelerate your Kubernetes clusters with Varnish Caching
 

Recognizing the cultural bias in Artificial Intelligence

Editor's Notes

  1. Hello. I wanted to give you an introduction to how I got involved with this topic. I am a student from Boise Idaho and my dream is to work in Silicon Valley. There are many stories and studies that come out saying Silicon Valley isn’t diverse but from where I come from this is an amazing place. At my school 1.6% of the student population is Black and it has an 80% White population. That makes for a lot of cultural bias in my classes, my school environment and the surrounding community. I got my first internship with HP two years ago in robotics and it opened up a new world for me. It gave me the opportunity to visit the Palo Alto office, learn more about the field which isn’t currently being taught at my school and ultimately accept an internship here. And what I learned was that cultural bias is everywhere, it is in our technology it is in our classrooms, and even the TV programs we watch. My first robotics project used computer vision. This was my idea and I had no idea where I was headed with this. But the idea was simple create a robotic hand and have it mimick my hand by using a 3D camera. This particular project didn’t prove to be particulary difficult but after a little research I happned upon a big problem, when it comes to facial recognition many people of color have difficulties being detected. I am always having difficulty using the snapchat filter on my face, I have to be in exactly the right lighting for it to work. So I set out to learn more about this technology I was working with everyday. [Cultural Bias almost prevented me from being an engineering student. I am not a first generation college student but I am a first generation engineering student. And with almost no foreknowledge of the type of classes I would take I jumped in knowing I had a passion for space science and engineering. But at school I encountered a couple of barriers, for example a place a I felt comfortable to study. You probably have never been to Boise State but we are not known for our study spaces. This is due in part because traditionally it is commuter school and in other part student cultures aren’t celebrated as much on campus so it can be hard to find a place you belong if you aren’t a part of Greek life. This led to me being really aware of the spaces I occupy. And I had to revaluate what I was at Boise State to do]
  2. I was driving with mu Uber driver to the airport one day and she asked what I did for a living. Flattered to be taken for a full fledged professional I told her about my work at HP. She was interested and we talked more. And it surprised her to realized that the tool Google Maps, which she was using to get to the airport, used machine learning. So where does machine learning and artificial intelligence affect our lives. I asked a few of my friends from twitter to tell me where they were using it, just to see how close my community was to the field….
  3. AI is popping up more and more, you may not realize how prevalent the field is and how near it is…Here are some more examples of everyday uses of AI…
  4. The implications: Word association reflects on your searches in Google, Amazon and LinkedIn LinkedIn Searches In LinkedIn searching for ‘Stephanie Williams’ would yield results for ‘Stephen Williams’ The fix: For Linked in the fix was to remove the ability to propose alternative names The implications: For male names and similar female name was not suggested. This suggests a bias in search results for male names over female names
  5. This is the story that really got me thinking about cultural bias. How could something as unchanging as the color of a person’s skin prevent them from being able to enjoy any product I create. And how would I handle a personal encounter with this issue? It’s not hard to imagine right? If you have a field that is not very diverse and your machine learning models are limited to people who look the same there will be gaps. But there are some really big cultural gaps that have occurred, for example what happens when your image recognition software misidentifies someone with darker skin as Google did?
  6. Maybe you see where I am going here. But in the past few months a lot of conversation has taken place about search and recommendation algorithms. Your facebook feed, your Twitter feed and more. We are starting to wonder how it affects our developing societal issues.
  7. Dylan Roof is one example. He is quoted as saying his Google search results for Black on White crime led him do what he did. When really the search algorithm based on reputability brought up various hate group websites with their own propaganda. In the case of our recent election it is less clear exactly how search algorithms played into this. But Facebook became a direct source and still is today for information and misinformation on American politics. Eli Pariser gave a Ted Talk that I highly suggest you watch the talks about filter bubbles. This talk is from 2011 and he shares one example of how search engines tailor your search results to what you see. One person searching on Google from their personal computer, in a certain location could see something entirely different. And this is what Facebook has also done in the past and Twitter for their trending news section. You location, your past use and searches are affecting what you see. So is it really a democratic, unbiased action we are taking when we search Google? No it’s not. Do we see everything our friends our seeing in the Trending section. Probably not.
  8. Facebook and LinkedIn have recently addressed both issues. LinkedIn was called out for having search terms that favor male names over women’s names. And with the unpopularity of the recent election and the up and coming idea of fake news, Facebook has also taken the same measure. Simple remove the preference that allows this unbalance to happen. This is similar to Google’s fix of simply removing the Gorilla tag from the tagging system. But Eli Pariser from the last slide recommended a couple of different things: Allow users to control and update what the algorithm does. And instead of basing results mainly on user preference also include in the search results items of importance or major news. Sprinkle in a couple things that ‘challenges’ the user. And provide some variety. This doesn’t remove the reamarkable new ability our technology has that over millions of users, multiple locations it can tell what we like and what we want to see. But in the age of information, and reliance on services like Google, it might be a new responsibility to provide a broad overview.
  9. I mentioned a few mistakes we have seen so far. And the cultural gap that is becoming visible in our age of new technology. But why does this gap occur?
  10. I believe it is based on who develops the technology and who uses the technology. And this is where my personal interest comes in. What can we be doing to diversify both pools. Like I said before it makes sense if you are White and you have White colleagues your testing set might not be diverse as it can be. Simillarily, if you look add a couple of dark skinned people to your data set it won’t help you match everyone because there are so many different kinds of brown. When it comes to our use of algorithms and how they associate our searches with our preferences, we need to be getting the most use from different areas. We need to break down the continued restrictions and barriers for communities to use the internet and access technology.
  11. There are a couple of emerging fields that I want to bring your attention too, that I think will provide an exciting edge to how we take care od this in the future. Not everyone jumps for joy at the sound of an ethics committee. But lets create technology that allows us to de-bias our algorithms.
  12. Explainable AI (XAI) is very cool to me. It asks why an algorithm has made a certain choice and if there is a bias it provides an opportunity to change that bias. For examples here is a street sign, and in the next column one or all of these segments is what the algorithm based its decision on. Then it makes the wrong or right guess and is attributed an accuracy. I saw a very similar demonstration last summer by a research and Washington State who was talking about Transparency in AI (another term to remember for the future). In his demonstration he was looking for an algorithm to detect wolves, and instead of the wolf in the picture the algorithm decided it was the snow that identified it is a wolf. This is what explainable AI seeks to do, understand why an algorithm has made its choice.
  13. Now during my search online I didn’t find a plethora of open and visible organizations that are working on this problem. But I am sure there are think tanks and other initiatives that I don’t have access too. If you know any feel free to let me know or tweet it at me later. But Mechanical Turks has seen some benefits in rooting out bias and then there is the Joy from Coded Gaze who has also launched an initiative called Algorithm Justice League. Tying this back to why I want to be involved. I mentioned that who develops this programs and who uses these programs is really important. This leads me to one of my favorite phrases representation matters. In the recent months we have seen a rise in available media online reaching out to women of color in tech. Women of color make up about 2% of the tech workforce. From my perspective as I work with algorithms and advanced technologies immediate problems arise that could be different from what someone else sees. And I want those problems addressed before the product goes on the shelf. With more technology we are going to have more word associated, user preference algorithms out there. We will have more computer vision and less controls.
  14. From my perspective as I work with algorithms and advanced technologies immediate problems arise that could be different from what someone else sees. And I want those problems addressed before the product goes on the shelf. With more technology we are going to have more word associated, user preference algorithms out there. We will have more computer vision and less controls. While I am in Boise I am hoping to spark conversations about this issue. And my goal is to bring down some of the hype of diversity that my community and maybe yours too flings around every chance they get. And bring those to real applications. I am continuing my work at HP as well with machine learning and computer vision and you can bet I will be there to pose the question so I noticed this about the robot when I used it, can make that better?