|
|
Clearview Faces Regulatory Scrutiny; Data Scarcity Hobbles Smart Cybersecurity
|
|
|
|
|
|
|
|
Welcome back. Too much data or not enough? Regulators, software developers and private companies are dealing with thorny issues around privacy, sparked by the expanded use of artificial intelligence in surveillance and enforcement. Facial recognition in particular has been cast in the national spotlight, as Federal Bureau of Investigations agents tap digital tools to track down suspects in last month’s Capitol riot, by scraping photos and videos posted online. Elsewhere, researchers say a scarcity of data is impeding efforts to leverage AI in fortifying cybersecurity from hackers, who are increasingly using the same technology to improve attacks.
|
|
|
|
|
|
|
|
Matthias Marx, top left, received these images when he asked Clearview AI what data they had on him. Mr. Marx pixellated some of the personally identifying information before providing it to the Journal. After he complained, the Hamburg data protection regulator ordered Clearview to remove his biometric data. The company has also said it would delete the data of any person in the European Union upon request.
PHOTO: MATTHIAS MARX
|
|
|
|
Clearview Tests Privacy Laws
Regulatory clampdowns in Europe, North America and Australia are forcing New York-based facial-recognition developer Clearview AI Inc. to tailor its services to different sets of regional privacy laws across the globe, The Wall Street Journal’s Catherine Stupp reports.
Hamburg order. Germany’s privacy authority recently issued a precedent-setting order to delete biometric data compiled by Clearview on a 32-year-old doctoral student (pictured above), under the European Union’s 2018 General Data Protection Regulation.
Pending investigations. Both the U.K. and Australian privacy officials are looking into similar issues, with a focus on the company’s use of biometric data and data scraped from the internet.
|
|
|
Rankling Canada’s Watchdogs
Canadian privacy regulators last week ruled that Clearview violated federal and provincial laws by collecting data on individuals, though the agency acknowledged having limited enforcement powers, The Wall Street Journal’s Jared Council reports.
|
|
|
Lack of consent. After a series of investigations, officials with four Canadian regulatory agencies said the company failed to get permission to use individuals’ biometric data, as required in Quebec, among other violations.
|
|
|
|
Lack of restrictions. Clearview attorney Doug Mitchell says the company only collects public information from the internet, which is explicitly permitted under Canada’s main privacy law.
|
|
|
|
|
|
|
3 Billion
|
|
The number of photos of individuals Clearview AI Inc. has scraped from the internet to search for matches using facial-recognition algorithms.
|
|
|
|
|
|
|
|
|
A law-enforcement officer looks for DNA samples on a t-shirt at the judicial district of the French national Gendarmerie near Paris. A New Jersey appeals court ruled that a defendant in a criminal case is entitled to obtain access to the source code and underlying documents of DNA-matching software used to identify him.
PHOTO: MARTIN BUREAU/AGENCE FRANCE-PRESSE/GETTY IMAGES
|
|
|
|
Prying Open 'Trade Secrets'
Courts and lawmakers are looking to curb the trade-secrets privileges granted to companies, which can be invoked in civil and criminal cases to protect intellectual property, The Wall Street Journal’s David Uberti reports.
Secret code. A New Jersey appeals court last week allowed a defendant in a criminal case to obtain the source code for the AI-powered software used by police to identify him.
A risky balance. Some executives and prosecutors say forcing such disclosures can open the door to IP theft or cyber threats.
|
|
|
|
|
|
|
“I think this is something that we really, really need to pay attention to—right now.”
|
|
— Intel research scientist Anna Trikalinou on the use of AI in cyberattacks.
|
|
|
|
|
|
|
Dearth of Data
Researchers and cyber specialists say the use of AI in fortifying cybersecurity is being hampered by a scarcity of data needed to train models, The Wall Street Journal’s James Rundle reports.
|
|
|
Slow start. Applications of AI within the U.S. government are still in exploratory phases, said Matt Hayden, the former Department of Homeland Security assistant secretary of cyber, infrastructure, risk and resilience.
|
|
|
|
Private sector progress. Intel Corp. says it’s working with Microsoft Corp. on using AI to analyze and classify malware variants by examining their coding.
|
|
|
|
|
Join us on March 31 for the WSJ Pro AI Forum, with panel discussions and interviews with corporate thought leaders, case studies, breakout groups and more. Topics will include reinventing the supply chain, AI in medicine and the revolution in tech regulation. More information is available here.
|
|
|
|
|
|
Engineers are increasingly using AI to streamline engineering work. A Rolls-Royce Pearl 700 engine ready for testing at the company's headquarters for business aviation near Berlin.
PHOTO: ROLLS-ROYCE
|
|
|
|
Rolls-Royce hits jet speed. Enginemaker Rolls-Royce is equipping some of its aerospace design engineers with AI tools in an effort to speed up the yearslong process of designing and testing jet engines. (The Wall Street Journal)
Wendy’s gets personal. Kevin Vasconi, the burger chain’s CIO, says he wants to develop AI-enabled apps that can learn to better predict customers’ taste from patterns in past orders, among other digital strategies. (The Wall Street Journal)
Study finds diagnostic bias. Research, published in Nature Medicine, found doctors tend to misgauge levels of pain reported by Black osteoarthritis patients by using an AI algorithm to compare more than 4,000 X-rays and with radiologists’ diagnoses. (BBC)
Economists: Robots threaten jobs. Raising the minimum wage could have the unintended consequence of accelerating the use of AI-enabled automation at big box stores, as business owners seek to cut labor costs, some economists warn. (Valley News)
Palantir trades higher on AI deal. Shares in the analytics firm rose 15% Monday after the company announced a partnership with IBM to develop AI tools for business. (Business Insider)
Software flags health risks. Sanford Health doctors are using an AI-powered analytics model to predict whether a person is at risk of developing self-harm, psychosis, addiction and other behavioral conditions. (Sanford Health)
Great Wall has views to Horizon. China-based auto maker Great Wall Motor is investing in auto chip startup Horizon Robotics and will jointly develop smart car technologies, including autonomous vehicles. (Reuters)
Auto maker gets driverless license. Vinfast, a unit of Vietnam’s largest conglomerate Vingroup JSC, says it has obtained a permit to test autonomous vehicles on public streets in California. (Reuters)
Earth Robots Invade Mars! Three robotic spacecraft this week are expected to land on Mars within days of each other, including the United Arab Emirates’ orbiter, China’s orbiter-rover combo and NASA’s rover, each one on a mission to collect and return rocks to Earth. (Associated Press)
|
|
|
|
|
Tesla Inc. said Monday that it bought $1.5 billion in bitcoin, a disclosure that follows Chief Executive Elon Musk’s promotion of the cryptocurrency and other digital-currency alternatives on Twitter. (The Wall Street Journal)
Apple Inc.’s talks with Hyundai Motor Group have broken down without an agreement for the South Korean auto giant to assemble vehicles for the iPhone company, Hyundai affiliates said Monday. (The Wall Street Journal)
WordPress VIP announced plans to buy content analytics company Parse.ly, parent company Automattic Inc.’s largest deal by cost and revenue. (The Wall Street Journal)
|
|
|
|
|
|
