This week in AI & Machine Learning: An expert discussion on AI’s role in our everyday lives, Google’s methods for teaching its chatbot, and more.
A Note from the Author
Energy providers are making strides to meet ambitious sustainability goals and transform their industry. Both malfunctioning equipment and ordinary wear and tear, however, still contribute to tens of thousands of tons of unintended volatile organic compound (VOC) emissions each year. Even organizations that have taken serious precautionary measures to address leakage could be missing out on warning signs and leaks without an AI-powered monitoring process. Learn how vision AI can help proactively detect leaks and support additional use cases for enterprises in the energy sector.
Artificial Intelligence News
Talking Points from the AI Commission’s Latest Field Hearing
The United States Chamber of Commerce’s Commission on Artificial Intelligence, Competitiveness, Inclusion, and Innovation (the AI Commission) held a field hearing in Palo Alto, California earlier this month to address closing the “trust gap” between employers who leverage AI solutions and employees who often worry they’re doomed to be replaced. Experts in attendance included executives, technologists, and members of Congress like California’s Anna Eshoo and Ro Khanna. Doug Bloch, Political Director for Teamsters Joint Council 7, suggested that the question is not whether workers should expect to be replaced (he doesn’t think they will), but how they can realize the maximum benefit of AI deployments. “We have to empower workers to not only question the role of technology,” he remarked, “but also to use tools such as collective bargaining to make sure that workers also benefit.” The AI Commission’s next hearing is scheduled to take place in London on June 13th. Check out a detailed summary of this recent hearing as well as recording and write-ups from past events
How Google Assitant Learns to Recognize Context
If we didn’t have the ability to draw conclusions from context clues, our conversations with one another would be a lot more repetitive and tedious. This is often the case when dealing with conversational AI, forcing users to carefully word their questions or wait on hold for a human to handle their queries. In a new blog post, researchers from Google describe how they deploy technology to help the Google Assistant easily gather context from past conversations. A common approach involves internally rephrasing the question so that it includes relevant additional information. Rephrasing effectively enables the Google Assistant to answer each question as a standalone query rather than relying on follow-ups. Check out the blog to learn more about how the Google Assistant interacts with users.
AI Can Often Predict Race from X-Rays, Experts Are Concerned
It’s not always a good thing when artificial intelligence recognizes something a human observer cannot – especially when nobody knows how. An international team of researchers, including scientists from Harvard and MIT, have found that AI can predict a patient’s race based on x-rays with around 90% accuracy. In a new paper published in The Lancet Digital Health, they explore the degree to which AI can discern race, attempt to determine how it makes these judgments, and offer examples of the problematic ways racial bias can manifest itself in AI-driven diagnosis. While the researchers have confirmed that AI is remarkably accurate at making these predictions (even when presented with degraded imagery), they are no closer to understanding how it makes its predictions. As such, they worry it may be impossible to create AI models for medical imaging that do not reflect at least some racial bias.
Join our Community
See you next week! Until then, keep the conversation going on Plainsight’s AI for All Slack Channel.
About the Author & Plainsight
Bennett Glace is a B2B technology content writer and cinephile from Philadelphia. He helps Plainsight in its mission to make vision AI accessible to entire enterprise teams.
Plainsight’s vision AI platform streamlines and optimizes the full computer vision lifecycle. From data annotation through deployment, customers can quickly create and successfully operationalize their own vision AI applications to solve highly diverse business challenges.
