Inside | Real news, curated by real humans
Inside AI

Inside AI (Jul 8th, 2019)

1. The FBI and Immigration and Customs Enforcement agencies have been using driver's license photos to feed data to thousands of facial-recognition searches. The agencies are scanning through millions of people's photos without their consent, according to a report in The Washington Post, which obtained data backing up the findings from Georgetown Law’s Center on Privacy and Technology. As The Post notes, neither Congress nor state legislatures have approved the system's development. The FBI has conducted more than 390,000 facial-recognition searches of federal and local databases since 2011, records show. - WAPO

For more stories like this, check out Inside Security, our newsletter with in-depth cybersecurity news and analysis.

  • Email gray
  • Permalink gray

2. Researchers at Lancaster University trained an AI to spot galaxy clusters in deep space. The “Deep-CEE” (Deep Learning for Galaxy Cluster Extraction and Evaluation) technique can identify and classify the clusters faster than a human using images of the cosmos. John Stott, a supervisor at the university, said the method is expected to locate thousands of clusters that haven't been seen before. You can read more about Deep-CEE online here. - GEEK.COM

  • Email gray
  • Permalink gray

3. More than three dozen call centers have started using the Cogito AI system in the last year, Time magazine reports. The program was trained to "hear" factors like tone, pitch, and word frequency in customer-service interactions. When it hears something off, such as irritation in a customer's voice, it notifies a worker to change their behavior, including trying to sound more sympathetic, slowing down their speech, or talking more. Cogito Inc. has raised more than $70 million from investors such as Goldman Sachs, OpenView Venture Partners, and Salesforce Ventures. Its clients include Humana and MetLife. - TIME

  • Email gray
  • Permalink gray

4. An AI-trained robot can automatically cut and pick iceberg lettuce in fields. Engineers from the University of Cambridge trained Vegebot using a machine-learning algorithm that helps it identify healthy heads of lettuce in fields under various weather conditions. The robot can also harvest the vegetable by cutting it at precise points. It averages about 32 seconds to pick each lettuce, which is still slower than a human, but researchers believe that will change as production materials become lighter. - NEW ATLAS

  • Email gray
  • Permalink gray

5. Providence St. Joseph Health, which has hospitals in seven U.S. states, signed on to use Microsoft's Azure and AI tools. The health-care company plans to use the technology to track health data, including cancer therapies and surgery outcomes, electronically. Other companies using AI in the medical field include Google, which is using AI to make better predictions about health, and Amazon, which is working with Boston's Beth Israel Deaconess Medical Center to examine how machine learning can simplify medical care. - BLOOMBERG

  • Email gray
  • Permalink gray

6. An AI system that taps Google Trends data can predict the severity and scope of flu outbreaks. A team from the University of Tokyo leveraged a sequence-to-sequence with attention model to improve the precision of the estimates, which is described in a paper, "Sequence to Sequence with Attention for Influenza Prevalence Prediction using Google Trends", on Arxiv.org. - VENTURE BEAT

  • Email gray
  • Permalink gray

7. A new report published by Vangavis singled out Bulgaria as a country with a high potential in the AI field. The Balkan nation has received boosts from a growing data science and software community as well as a lot of access to venture funding. In Bulgaria, nearly 30 percent of companies surveyed for the report were established over the last two years, indicating a startup wave in the AI market. - EMERGING EUROPE

  • Email gray
  • Permalink gray

8. British online supermarket Ocado has revealed that it's planning to open a technology hub in London that will create 260 jobs focused on AI, cloud, and data. According to the company, the hub will initially support 40 software engineers; however, the aim is to increase that number to 300 "in the near future," with 260 of those positions related to technology roles. - IT PRO

  • Email gray
  • Permalink gray

9. Only a quarter of organizations polled in a recent survey said they have a broad strategy for AI projects. The IDC, which conducted the global survey of 2,473 organizations, also found that lack of workers, costs, and biases were their reported roadblocks to AI implementation. - ZDNET

  • Email gray
  • Permalink gray

10. A machine-learning system was trained to identify which members of the Beatles wrote certain songs. Scientists from Harvard and Dalhousie University in Canada trained the system using 70 songs, which allowed it to build a "musical fingerprint" and attribute song writing to either John Lennon or Paul McCartney. In the songs "Ask Me Why," "Do You Want to Know a Secret" and "A Hard Day's Night," the system was able to identify the songwriter with a 90 percent certainty. Their findings were published in Harvard Data Science Review. - THE INQUIRER

  • Email gray
  • Permalink gray

Written and curated by Beth Duckett in Orange County. Beth is a former reporter for The Arizona Republic who has written for USA Today, Get Out magazine and other publications. Follow her tweets about breaking news and other topics in southern California here.

Editing team: Kim Lyons (Pittsburgh-based journalist and managing editor at Inside), David Stegon (senior editor at Inside, whose reporting experience includes cryptocurrency and technology), and Bobby Cherry (senior editor at Inside, who’s always on social media).

  • Email gray
  • Permalink gray

Subscribe to Inside AI