Embedded Files
Building an ethical framework for AI
Building an ethical framework for AI
February 26, 2018 4:24 PM
Design #1
Zeynin Juma at 2018-05-03 9:39 PM
PLAN
GOALS
S: Design ethical framework
M: When the HOW, WHAT, WHEN, WHERE, WHY are defined
A:
R:
T: By June 1, 2018
S: Design emotional intelligence library
M: When the HOW, WHAT, WHEN, WHERE, WHY, WHO are defined
A:
R:
T: By June 30, 2018
S: Design penny transaction framework
M: When the HOW, WHEN are defined
A:
R:
T: By August 31, 2018
Problems
July 1, 2018
11:26 AM
Zeynin Juma at 6/7/2018 10:00 PM
Program Emotional Intelligence.
What problem am I solving?
Human to computer interaction is cold. Machines are not emotional beings, but they can be more aware of their human counterparts.
Why is this important to me?
I've spent so much time and effort on personal growth and I want to leverage my knowledge to help others. Plus, I'm a programmer, not a hardware engineer. AI and ML or building a library is my best option.
Couldn't I just build a protocol for human to machine communication?
SWAT
SWAT
Zeynin Juma at 2018-06-29 8:45 PM
PROBLEM: when chatting with 'bots, how do you know when your customer feels frustrated and requires human intervention?
SOLUTION: a penny for your thoughts or feelingz.io will provide you with the emotional intelligence you need to ensure your customers stay happy and keep coming back. Consistent customer service provides a series of scripts to keep each engagement fresh, but still provides a unified message that uniquely expresses your brand.
KEY METRICS: Number of pennies amassed, adoption rate
UNIQUE VALUE PROPOSITION: No one else can tell you how your client is feeling based on a chat conversation. This cost-effective solution can drive store-front automation, while signalling when best to engage a client with another human, thus freeing up time to complete more important tasks.
UNFAIR ADVANTAGE: tapping into the harmonic frequency of the earth; using 10 Hz signal superimposed on an emotional score will input into my model. I'm a female developer who has spent the better part of my thirties learning about emotional health.
PLAN
Pay for coursera's ML
Read Andrew Ng's book
Is the way forward through emotions? Can I create a model for emotions? Can I create a model for ethics? How do you model behaviours vs. actions?
Study business ethics
Write about your trials, tribulations and trepidations on LinkedIn
Ethical Framework Requirements:
Criteria for usage: if this is AI, run all decisions through this library.
Input from as many sensors as available. What is the minimum?
Returns a judgement and prescribed actions.
Output object of entire emotional profile (empathy package) + learnings, for future storage.
DESCRIPTION
Humans spend a life-time building beliefs and morals that shape every decision they make. Artificial Intelligence inherently lacks this belief system and cannot conscionably decide what is "right" and what is "wrong". Furthermore, who is this judgement relative to? Humans or machines?
The Ethical Framework provides this backdrop upon which all decisions are made. With strong ethics, artificial beings have a context from which to decide what is in the best interests of humans, and machines, but with a clear hierarchy that biological life is priceless.
It is naïve to think that one woman can conquer all this. This is a starting point. The ultimate goal is to establish a world consortium that can reveal issues and drive real solutions and help build a world standard (like GSM).
EMPATHY
Measured responses use empathy to create a relative moment in time from which to measure ethical behaviour. Empathy differs from emulation as the emotion extrapolated from the sensor data, is used to determine relative ethical conduct and reactions are then appropriately described.
Empathy is a derived moment of emotion, taken from human interaction. Derived human emotion, plus situational information along with ethical context provide information to help determine a reaction.
RESEARCH
What input will be required to make an emotional decision?
How will output be delivered – like-for-like from input? To be defined
INPUT
All present-state sensor information. Will this need to be tracked over time? A cubed view will help with decision making.
OUTPUT
Defining the output is much more complex as each AI form-factor is different. The assumption is that physical reaction/actions are provided, but the output may be (but not limited to) verbal, appendage movement, expression of facial features, direction, displacement, or text.
Ethically Aligned Design, First Edition
Ethically Aligned Design, First Edition
https://standards.ieee.org/content/dam/ieee-standards/standards/web/documents/other/ead1e.pdf
If you are not already involved in The IEEE Global Initiative on Ethics of Autonomous and Intelligent Systems, we invite you to join the open community of over 2,000 global experts focused on cutting edge ethical and values-driven issues in autonomous and intelligent systems and in moving principles to practical application. Join to receive our regular newsletter, announcements, and to get an invitation into our private EAD community.
The Mission and Results of The IEEE Global Initiative on Ethics of Autonomous and Intelligent Systems is to ensure every stakeholder involved in the design and development of autonomous and intelligent systems is educated, trained, and empowered to prioritize ethical considerations so that these technologies are advanced for the benefit of humanity.
To advance toward this goal, The IEEE Global Initiative on Ethics of Autonomous and Intelligent Systems brought together more than a thousand participants from six continents who are thought leaders from academia, industry, civil society, policy, and government in the related technical and humanistic disciplines to identify and find consensus on timely issues surrounding autonomous and intelligent systems.
By “stakeholder” we mean anyone involved in the research, design, manufacture, or messaging around intelligent and autonomous systems—including universities, organizations, governments, and corporations—all of which are making these technologies a reality for society.
JOIN
IEEE, 445 Hoes Lane, Piscataway, NJ 08854-4141 USA© 2019 Copyright Contact Privacy Policy
Affiliates
May 1, 2019 3:11 PM
https://www.ualberta.ca/kule-institute/news-events/ai-ethics-and-society-conference
International Review of Information Ethics (IRIE): http://www.i-r-i-e.net/
International Review of Information Ethics
<<Top 10 Strategic Technology Trends for 2019.one>>
Submissions: https://standards.ieee.org/news/2016/ethically_aligned_design.html
READ: https://standards.ieee.org/industry-connections/ec/ead-v1.html
https://standards.ieee.org/industry-connections/index.html
https://standards.ieee.org/industry-connections/ec/news.html
From: IEEE Standards Association <ieeestandards@ieee.org>
Sent: May 2, 2019 4:31 PM
Subject: As requested: Ethically Aligned Design, First Edition
Thank you for your interest in
Ethically Aligned Design,
First Edition: A Vision for Prioritizing Human
Well-being with Autonomous and Intelligent Systems
If you are not already involved in The IEEE Global Initiative on Ethics of Autonomous and Intelligent Systems, we invite you to join the open community of over 2,000 global experts focused on cutting edge ethical and values-driven issues in autonomous and intelligent systems and in moving principles to practical application. Join to receive our regular newsletter, announcements, and to get an invitation into our private EAD community.
The Mission and Results of The IEEE Global Initiative on Ethics of Autonomous and Intelligent Systems is to ensure every stakeholder involved in the design and development of autonomous and intelligent systems is educated, trained, and empowered to prioritize ethical considerations so that these technologies are advanced for the benefit of humanity.
To advance toward this goal, The IEEE Global Initiative on Ethics of Autonomous and Intelligent Systems brought together more than a thousand participants from six continents who are thought leaders from academia, industry, civil society, policy, and government in the related technical and humanistic disciplines to identify and find consensus on timely issues surrounding autonomous and intelligent systems.
By “stakeholder” we mean anyone involved in the research, design, manufacture, or messaging around intelligent and autonomous systems—including universities, organizations, governments, and corporations—all of which are making these technologies a reality for society.
LEARN: https://standards.ieee.org/industry-connections/ecpais.html
Overview Document: https://standards.ieee.org/content/dam/ieee-standards/standards/web/documents/other/ead1e-overview.pdf
Videos and Webinars: https://standards.ieee.org/industry-connections/ec/resources.html
Bios: https://standards.ieee.org/content/dam/ieee-standards/standards/web/documents/other/ec_bios.pdf
TechEthics: https://techethics.ieee.org/
Beyond: https://beyondstandards.ieee.org/
How to Participate
How to Participate
To learn more about the Initiative, please contact one or more of the following:
John C. Havens, Executive Director of The IEEE Global Initiative (John Havens)
The Industry Connections Program Administrator
Chair: Raja Chatila, CNRS-CNRS-Sorbonne UPMC Institute of Intelligent Systems and Robotics, Paris, France; Member of the French Commission on the Ethics of Digital Sciences and Technologies CERNA; Past President of IEEE Robotics and Automation Society (Raja Chatila)
Innovation and Responsible AI
March 13, 2021 11:42 AM
A talk by Yina Moe-Lange
Product Manager | Author, 2021.AI | The Big Y
From <https://yyc-datacon.heysummit.com/talks/innovation-and-responsible-ai/?ah=PvLHbr5b>
Accounting framework
THINGS TO CONSIDER:
Google examples
Before getting started, take some time to think about the project: who impacted, where data comes from, what are you getting into. Consider what can go wrong. What representation is - negative extranalities?
What is considered responsible in your organization? For yourself? What are your principles when analyzing data?
Page updated
Google Sites
Report abuse