AI is a powerful technology capable of making enormous contributions to our economy and society. However, it is essential that the diverse users of AI are equally reflected by those charged with building this important technology. Raia Hadsell, a research scientist at Google DeepMind, noted “responsible AI is a job for all” during the Lesbians Who Tech Pride Summit last week. Hadsell emphasizes that regulators, lawyers, ethicists, and philosophers play a critical role, but researchers and scientists can actively build responsible AI.
When considering research approaches, bias can emerge anywhere in the system, with race and gender the least obvious biases to detect. In this thread, Mozilla Fellow Deborah Raji gives an overview of some misconceptions about race and gender biases in algorithms. More specifically, sexual orientation and gender identity are unobservable characteristics, breaking the assumptions of many algorithmic fairness approaches. Researchers and activists have concerns around privacy, censorship, language, online safety, health, and employment for people who identify as LGBTQIA+ when AI is applied without appropriate consideration.
That being said, AI has also been used to support human rights awareness projects. Last year a team of researchers behind the award-winning film Welcome to Chechnya used the same AI technology seen in deep-fake videos to overlay volunteers’ faces on people telling their survival stories. LGBTQIA+ people in Chechnya have faced significant persecution including unlawful detention, torture, and other forms of abuse. The tech helps to better display the survivors’ emotions, shining a light on human rights abuses while minimizing the risk for victims involved in the production.
To support the people behind AI technology, organizations such as Queer in AI work to create awareness of queer issues in AI/ML, foster a community of queer researchers and celebrate the work of queer scientists. Visit Queer in AI to view their resources and learn more.
AI News This Week
-
Chips off the old block: A new batch of semiconductor startups hopes to restore Canada’s hardware edge – Subscription required (The Logic)
Semiconductor shortages have idled factories globally, and yet Canada’s semiconductor sector is experiencing an AI-inspired revival. Toronto AI chip companies, such as Radical Ventures portfolio company Untether AI, have a talent advantage thanks to the city’s key role in advancing AI technologies. Radical Ventures co-founder Tomi Poutanen notes success for semiconductor players rests on the ability to build complementary software and points to the Toronto ecosystem as a strong competitor in this field. Radical Ventures Partner Salim Teja, who is a founding member of Canada’s Semiconductor Council, suggests chips designed for machine learning, edge or distributed computing, and quantum applications could be Canadian areas of specialization. Salim also notes that a national chip policy is still needed to support Canada’s AI strategy with the necessary hardware.
-
These are the startups applying AI to tackle climate change (Forbes)
AI can reduce climate change based risks and prevent further disaster. This article outlines some startups doing this work, including ClimateAI. These ‘climate intelligence’ startups use prediction to help organizations anticipate and prepare for extreme weather events. This is achieved by combining machine learning with traditional weather modeling techniques to unlock publicly available, proprietary, and customer-specific data to train their models.
-
Video: Workshop on Autonomous Driving keynote for the conference of Computer Vision and Pattern Recognition (CVPR’21 WAD)
Raquel Urtasun, founder of Radical Ventures portfolio company Waabi, gave the keynote for the Workshop on Autonomous Driving (WAD) at CVPR’21 – the premier annual computer vision event comprising the main conference and several co-located workshops and short courses. The Waabi team has two papers that have been selected as Best Paper finalists.
-
Walmart’s AI is getting smarter about grocery delivery (TechCrunch)
Applying AI to grocery delivery is not new in 2021. But, Walmart is leveraging the data of over 200 million people who shop at Walmart in-store and online each week for more than 150,000 different grocery products to predict substitution preferences based on consumer behaviour. So far, customer substitution acceptance rates have increased.
-
Rembrandt’s damaged masterpiece is whole again, with AI’s help (The Verge)
Rembrandt’s The Night Watch was unceremoniously cropped to fit a small wall in the 18th century. Today, Rembrandt’s style can be imitated by an AI after two-years of training on scans, x-rays, and 528 digital exposures of the original painting. The AI is being used to fill in the blanks, helping viewers imagine how the original work might have looked when first completed.
Radical Reads is edited by Leah Morris (Senior Director, Velocity Program, Radical Ventures).