0:0013:13

Changing the World Through AI and Training Data

Wendy Gonzalez 13:13 2020

The AI training data behind self-driving cars can also be a way out of poverty

Before a self-driving car can tell a bicycle from a motorcycle, people must label the data it learns from. In Kenya's Mathare, where youth and women unemployment hovers at 70% or more, that labelling became a first real job. Wendy Gonzalez, President and Interim CEO of Samasource, which serves 25% of the Fortune 50, shows what happens when companies give work instead of aid, and whether it lasts.

See how Samasource's hiring model, quality approach and impact data fit together

Watch on YouTube

In short

Wendy Gonzalez, President and Interim CEO of Samasource, explains what AI training data is and why high-quality labelling decides whether systems such as self-driving cars work. She describes Samasource's model of hiring people living below the poverty line in Kenya and Uganda, training them in data annotation and paying living wages, and shares impact results, founder Leila Janah's legacy and her experience raising funding as a minority woman executive.

“And our belief inherently is that talent is distributed equally while opportunity is not.”
“Imagine a self driving car that can detect a motorcycle, but not a bicycle or can read certain road signs, but not others.”

What you will learn

  • Explain what training data is and why AI models cannot exist without it.
  • Recognize why last-mile labelling quality matters for safety-critical systems such as self-driving cars.
  • Describe an impact hiring model that recruits below the poverty line, trains people and pays living wages.
  • Measure social impact by tracking former workers' earnings years after they leave.
  • Frame diversity as a strength when seeking investors as a woman-led company.

Key takeaways from the speaker

  • TBA

Details

Level
Beginner
Format
Talk
Technical depth
Introductory technical
For
All career stages
Length
13:13
Published
July 2020

Questions this talk answers

What is AI training data?

Before a machine can learn to see, hear or speak it must be trained on labelled data. A self-driving car, for example, must be taught to recognize vehicles, pedestrians, lanes and road signs.

Why does the quality of training data matter so much?

A self-driving car that detects a motorcycle but not a bicycle, or reads some road signs but not others, is a real risk. Getting the last mile of quality is critical.

How does Samasource decide whom to hire?

The only criterion is living below the World Bank poverty line, not education or job experience. Hires are trained in training data and AI, then employed full time with benefits and living wages.

How does Samasource know its model works?

It surveys workers who leave: three years later they earn on average four times their previous earnings, and well over 80% earn more than at Samasource or return to university.

What was it like to raise funding as minority women executives?

Gonzalez and Leila Janah were often the only women in the room, but they found investors who understood AI and also saw the value in their diversity, raising a $14.8 million Series A in 2019.

About this talk

In this session, President and interim CEO of Samasource, Wendy Gonzalez, will take a deep dive into the ways AI and machine learning are currently improving society, and how those technologies can be used to positively impact and change future outcomes -- from ending poverty to wildlife conservation.

AI has the power to change the world as we know it. We’ve already witnessed AI being used within the healthcare industry to power chatbots that help citizens get accurate health information in real-time. Right now, it’s helping scientists diagnose and find vaccines and treatments for COVID-19. Within the education industry, textbooks are being digitized with the help of AI - ensuring that our kids keep learning during stay-at-home orders. However, behind every successful AI platform is high-quality training data. There is no AI without it, as training data teaches AI to see, speak and think, making all of these innovations work.

But has AI reached its full potential? Where is there room to grow and how can we best utilize the technology to make the world a better place?

Since Samasource’s inception, social good has been at the core of everything we do, and as a result, we’ve lifted more than 50,000 people out of poverty in countries like Africa and Uganda, by equipping them with digital and machine learning skills to offer full-time employment and living wages. Through their work, we’ve powered AI for many groundbreaking advancements like those mentioned above and will use this session to speak to how machine learning and training data can be used to create meaningful technological advancements that shape our future.

I’ll dive into various use cases for AI and how we can ensure tech is being used to its full potential as we enter a new world impacted by a powerful and damaging global pandemic.

Topics

Related talks

Browse the library