Inside AI - February 26th, 2020 |

Inside AI (Feb 26th, 2020)

Clearview AI's client list breached / White House CTO talks AI R&D / Tweet: Bias in AI still a concern

Subscribe to Inside AI

Inside AI. ###PIXEL:IMG###
Subscribe | View in browser

1. Clearview AI, the company that created a facial recognition database for law enforcement, confirmed that its client list was stolen in a data breach. The Daily Beast first obtained a memo that was sent to the company's customers; the memo reads that an unidentified intruder gained “unauthorized access” to Clearview's entire client list, which is made up of federal and state law enforcement agencies that pay to access Clearview's restricted database and face recognition tech. The breach, which did not involve the more than 3 billion photos scraped from social media for the database, also accessed each client's number of user accounts and their number of searches. The company - which first gained notoriety when the New York Times published a story about its secretive nature and privacy implications - said its servers weren't breached and there its network and systems weren't compromised. "We patched the flaw, and continue to work to strengthen our security," Clearview said in a statement. - ARS TECHNICA

Why it matters: As today's Inside Social newsletter notes, the breach has the potential to increase anxieties "over the privacy-violating monster this company has created." Law enforcement agencies who utilize its database might be worried about the company's security flaws, especially since some may not want the public to know that they use it.

2. In related news, people using the social aggregator website Reddit started a discussion on the legality of scraping the web for data, which could be used in some machine learning applications. The original poster cited a recent ruling by a California appeals court, which found that it's generally legal to scrape public websites without approval to glean massive amounts of data; the ruling was based on LinkedIn's legal dispute with HiQ, a data analytics firm, which LinkedIn tried to (unsuccessfully) block from scraping its site. As of Wednesday, the Reddit post had nearly 50 comments. One user asked, "So does this mean I can repost the tinder dataset I had on Kaggle without getting a DMCA from Match group llc?" A responder noted that "You can scrape it, you can't share it legally." Another poster wrote that scraping "doesn't mean you're free to use" the data, citing an example of taking images from Disney's website and then sharing it, which they surmise would result in immediate lawsuits. - R/MACHINELEARNING / REDDIT

3. E-cigarette company Juul could be working on an AI-powered device that helps users quit nicotine by learning their vaping habits over time, according to a patent application that was recently made public. The patent, titled "Devices and Methods for Cessation of Nicotine Addiction," describes a device that would work alongside a vaporizer to slowly lower a person's nicotine intake, alternating between nicotine and non-nicotine products when they inhale. The patent notes that the controller would utilize machine learning to adjust which product is delivered, based on the users' previous habits. That data could also be shared with the user's health care provider, friends, or other third parties to help them quit, it says. Juul, which applied for the patent last summer, is facing lawsuits and criticism over its alleged deceptive advertising, which critics say has contributed to a rise in young people vaping. - THE LOGIC

4. U.K. startup Babylon Health, which developed a chatbot to guide patients experiencing health issues, has publicly criticized a doctor who brought up concerns about its AI-powered bot. Through his Twitter handle, Dr. David Watkins has cited cases in which the chatbot may have given patients poor advice regarding their health. In one alleged example, a 48-year-old smoker was experiencing central chest pains and sweating, and the app told him to see a general practioner rather than immediately call emergency services. Earlier this week, Babylon appeared to fire back against the criticism, saying that an anonymous Twitter user going by the name of @DrMurphy11 has "trolled" the company and targeted staff, clients and others while publishing defamatory content about the company. The AI bot has been used "millions of times, and not one single patient has reported any harm," according to Babylon. The company has asked @DrMurphy11 - who it did not identify by name - to publish his analysis of the AI and let it "be assessed by any objective expert." - MOBI HEALTH NEWS

5. In a new piece for Geekwire, White House CTO Michael Kratsios explains how the federal government's American AI Initiative will result in R&D efforts, which are sure to trickle down to Big Tech and AI research institutions in Seattle and elsewhere. According to Kratsios, the government's non-defense research spending for AI is proposed to be nearly $1 billion in the next budget cycle. By 2022, a two-year plan places it at $2 billion. “We think this is a game-changer moment for AI R&D in this country,” Kratsios said, noting that “People at the University of Washington and other institutions around the country will be able to apply and tap into those new funds." - GEEKWIRE

6. Google and 51 other healthcare and tech organizations, including the American Medical Association and IBM, released the first American National Standards Institute-accredited standard for the use of artificial intelligence technology in healthcare. The group, convened by the Consumer Technology Association, issues a framework that clearly defines terms that spark debate, such as remote patient monitoring and assistive intelligence, and more. "As health systems and providers use AI tools such as machine learning to diagnose, treat and manage disease, there's an urgent need to understand and agree on AI concepts for consistent use. This standard does exactly that," Rene Quashie, vice president of policy and regulatory affairs in CTA's digital health group, said in a statement. - BECKER'S HOSPITAL REVIEW

A version of this story first appeared in Inside Google.

7. ABB, an industrial robotics manufacturer for automakers, has teamed up with Covariant, an AI startup, to develop warehouse robots that can pick up a variety of objects better than the current ones out there. Covariant - which was founded by OpenAI researchers and those from the University of California at Berkeley - used reinforcement learning to create its robotic gripping software, which could help Switzerland-based ABB expand into the logistics industry. Following a contest put on by ABB last year, Covariant was selected because it was the only participating company whose software was capable of recognizing a variety of items without human help. - FORTUNE

8. Recommended Read: The Wall Street Journal published a story today detailing how governments are using machine learning and data analytics to try and find tax cheaters. The Journal notes that machine learning could detect more anomalies in tax filings, potentially raising revenues for governments in the long run. The AI systems can also reportedly help public agencies "respond to taxpayers’ questions and make themselves more efficient," it notes. - WSJ

9. By the numbers: Here is a list of AI startups that have raised venture capital funding in recent weeks:

  • (China), autonomous vehicles: $400M led by Toyota Motor Corporation.
  • SambaNova Systems (Palo Alto, CA), AI/data-sensitive app running platform from datacenter to the edge: $250M C led by BlackRock, participation from GV, Intel Capital, Walden International, WRVI Capital, Redline Capital.
  • Revl (San Francisco, CA), AI-based video editing company: $5M A led by Nimble Ventures, participation from Tuesday Capital, Silicon Valley Data Capital, Luma Pictures.
  • (Cambridge, UK), AI-based internet deep search platform for enterprise knowledge: $3.9M A led by Nauta Capital, participation from Cambridge Angels.
  • Chattermill (London, UK), AI-based customer feedback/understanding platform: $8M A led by DN Capital, participation from Ventech, btov Partners, Silicon Valley Bank, Matt Price, Nilan Peiris.

For daily updates in startup funding, consider signing up for our Inside Venture Capital newsletter.

10. Tweet of the Day: Blademy founder Ayoola John-Muyiwa let his followers know that bias in AI remains a "real concern." John-Muyiwa, who is also the creative director at Astro Media Lab, tweeted that "People of color must have a seat at the table when AI systems of the future are being developed." Not adequately addressing these biases could lead to risks, he notes, such as people of color being struck by driverless cars, receiving credit limits that are lower than their peers, and "humiliation." - @AYOOLAJOHN_ / TWITTER

Written by Beth Duckett, a former reporter for The Arizona Republic who has written for a variety of publications including USA Today, Get Out, and Images Arizona magazine. You can follow her tweets here.

Edited by Inside staff writer Elizabeth Barr.

Copyright © 2021, All rights reserved.

Our mailing address is:
767 Bryant St. #203
San Francisco, CA 94107

Did someone forward this email to you? Head over to to get your very own free subscription!

You received this email because you subscribed to Inside AI. Click here to unsubscribe from Inside AI list or manage your subscriptions.

Subscribe to Inside AI