A big data company held a beauty contest judged by Artificial Intelligence.
The evaluation factors included…
1. Wrinkles 2. Skin color and conditions 3. Difference Between Real Age and Perceived Age 4. Similarity to the model (100 people of all races and ages) 5. Face symmetry
Of the 6,000 participants, 44 were selected as the winners. However, many were dismayed by the results.
Out of 44 winners,
most of them were white,
five were Asians,
and only one was black.
What if AI dominates humans?
Artificial intelligence is already guiding many decisions for human beings.
Navigation apps to find directions,
algorithms that provide user preference content,
investment apps that provide good stock information.
Now humans just need to choose from several alternatives that AI offers.
Today’s artificial intelligence learns based on countless data created by humans. If the data is stained with human prejudice and discrimination, it is natural that the ‘beautiful human’ selected by artificial intelligence reflects it.
Not to Repeat the Mistakes of Mankind
Various ethical issues as well as ‘discrimination’ and ‘prejudice’ are being revealed through artificial intelligence.
The person who felt this problem most seriously was Yoshua Benjio, the grandmaster of deep learning.
There were two main causes of the problem, he thought.
1. Indiscriminate development of technology without ethical awareness
2. Competition-oriented market that prevents ethical considerations
To overcome them, Yoshua establishes Mila, an ethical AI development agency.
There are two principles of Mila.
1. Advancing artificial intelligence for the benefit of all.
2. Conducting social participatory research
Yoshua’s first move was to bring together people of various majors. He has gathered not only computer and AI majors, but also various majors such as ethics, philosophy, humanities, and sociology. What Yoshua suggested to them is a multi-faceted problem-solving method viewed by various majors. And the development of artificial intelligence with problem-solving methods.
Yoshua didn’t just look for social problems on the Internet and in books. As such, he has established various partnerships with stakeholders including public, private and corporate entities. It made researchers find problems in society by communicating with various stakeholders in society.
Mila’s “AI for the Benefit of All”
AI to eradicate human trafficking in Canada
AI deciphering programs with gender and racial bias
AI that Drives Drug Development for Treatments
AI showing what it will be like after climate change in 3D
AI to track safe location of COVID-19 patients
Opening the Age of AI Development for All
Social issues learned from various stakeholders
Diverse perspectives of multi-majors
Two factors made technological development successful considering ethical awareness.
And the positive interactions that Mila created have made Montreal a world-class innovative city.
And Yoshua reveals their success story to the world.
Montreal Declaration for the Responsible Development of Artificial Intelligence(10 principles) Well-being Respect for autonomy Protection of privacy and intimacy Responsibility Democratic participation Equity Solidarity Diversity Inclusion Prudence Sustainable development
These 10 principles are a concrete blueprint for mankind to develop with AI. In the midst of fierce competition, it is not easy to follow sound principles. Therefore, innovative AI technology that understands the problems of stakeholders has greater value.
* The Montreal Declaration is a set of AI development guidelines created by citizens, experts, public policymakers, industry stakeholders, civil society organizations, and experts under the theme of ‘Development of AI’s Social Responsibility.
Yoshua, the father of deep learning
“AI for the benefit of all”
This simple phrase shows Yoshua’s heartfelt wish that the technology he gave birth to will not be a monster but a savior of mankind.