Alibaba Seeks to Distance Itself from Uighur Facial-Recognition Software

A man stands near the logo of Alibaba at the company's office in Kuala Lumpur, Malaysia. Reuters file photo
A man stands near the logo of Alibaba at the company's office in Kuala Lumpur, Malaysia. Reuters file photo
TT

Alibaba Seeks to Distance Itself from Uighur Facial-Recognition Software

A man stands near the logo of Alibaba at the company's office in Kuala Lumpur, Malaysia. Reuters file photo
A man stands near the logo of Alibaba at the company's office in Kuala Lumpur, Malaysia. Reuters file photo

Chinese tech giant Alibaba has sought to distance itself from a face-recognition software feature devised by its cloud computing unit that could help users to identify members of the country's Muslim Uighur minority.

A report this week revealing the software feature made Alibaba, one of the world's most valuable companies, the latest Chinese corporate entity embroiled in the controversy over China's treatment of Uighurs.

In a statement posted online late Thursday, Alibaba said it was "dismayed to learn" that Alibaba Cloud developed the feature.

The technology was used only in for capability-testing and not deployed by any customer, Alibaba said, adding that it had "eliminated any ethnic tag" in its products.

"We do not and will not permit our technology to be used to target or identify specific ethnic groups," it said.

The Uighur issue looms as a worrying threat for Chinese companies as global criticism grows over Beijing's policies in the northwest region of Xinjiang.

Rights groups say as many as one million Uighurs and other mostly Muslim minorities have been held in internment camps there.

Beijing initially denied the camps' existence but now calls them vocational training centers.

Surveillance spending in Xinjiang has risen sharply in recent years, with facial recognition and other technologies deployed across the province.

Washington last year blacklisted eight Chinese tech firms for alleged links to the surveillance effort.

Last week, US-based surveillance research firm IPVM said Chinese telecoms company Huawei had been involved in testing facial-recognition software that could send alerts to police when Uighur faces were recognized.

Huawei denied the claim.



South Korea Team Showcases Drone That Serves as ‘Flying Shopping Cart’

 Park Hyun-geun, a graduate student of Mechanical Design and Robot Engineering at Seoul National University of Science and Technology (Seoul Tech), demonstrates the aerial transportation platform "Palletrone" on stairs at the Seoul Tech in Seoul, South Korea, October 10, 2024. (Reuters)
Park Hyun-geun, a graduate student of Mechanical Design and Robot Engineering at Seoul National University of Science and Technology (Seoul Tech), demonstrates the aerial transportation platform "Palletrone" on stairs at the Seoul Tech in Seoul, South Korea, October 10, 2024. (Reuters)
TT

South Korea Team Showcases Drone That Serves as ‘Flying Shopping Cart’

 Park Hyun-geun, a graduate student of Mechanical Design and Robot Engineering at Seoul National University of Science and Technology (Seoul Tech), demonstrates the aerial transportation platform "Palletrone" on stairs at the Seoul Tech in Seoul, South Korea, October 10, 2024. (Reuters)
Park Hyun-geun, a graduate student of Mechanical Design and Robot Engineering at Seoul National University of Science and Technology (Seoul Tech), demonstrates the aerial transportation platform "Palletrone" on stairs at the Seoul Tech in Seoul, South Korea, October 10, 2024. (Reuters)

South Korean researchers have developed a transport drone flying on multiple flexible rotors that self-correct to stay level in flight and can be used as a "flying shopping cart" to carry goods over uneven terrain such as stairs.

The prototype developed by a Seoul National University of Science and Technology team has a cargo platform mounted on top of a multi-rotor drone and is handled by a person using gentle force to guide the hovering aircraft.

Members of the team demonstrated the hovering platform with a handlebar much like one on a push shopping cart moving objects up and down stairs and loading boxes on top as it hovered mid-air and maintained its balance by using a center of mass estimation algorithm.

To move objects over uneven terrain or stairs when a wheeled cart cannot, the drone responds to human control with what the developers call a physical human-robot interaction technique that anticipates human intentions for smooth flight, said Lee Seung-jae, professor of mechanical system design engineering.

But the broader focus of Lee's team is not on developing a shopping cart to be used over steps, but instead on applications that would use a drone with reliable horizontal stability without pitching and rolling.

"The Palletrone can be more than a flying shopping cart," he said, referring to the name the team gave the prototype by joining the words pallet, which is the platform for cargo on top, and drone.

Lee's team has tested a platform to carry objects up to 3 kg (6.6 lbs), and concedes commercial applications for cargo transportation at such a small weight easily carried by humans is limited.

Still, the mechanism that allows the drone to change directions in flight without banking and to maintain level attitude has applications for delivering sensitive or fragile payloads, Lee said.

But Lee's team is looking further ahead for the technology's potential use for uncrewed "flying taxis" carrying humans and for drones to be "refueled" mid-air, by changing batteries so the aircraft do not need to return to base for a fresh charge.

Multi-rotor drones are inherently limited in speed and range compared to fixed-wing drones but have better control and maneuverability, including the ability to hover in flight.

They have been used to deliver cargo, food and medical supplies, but commercial applications have been limited largely because it is impractical to increase the size of the battery enough to transport a heavier payload over a longer distance.

Seoul Tech's work was published in IEEE Robotics and Automation Letters this year, the publication of the Institute of Electrical and Electronics Engineers based in New York.