Sun, Aug 25, 2024 page3

• China-made AI apps a concern: experts

  GROWING TREND: AI services provided by Chinese companies or funded by Chinese investors are a security risk, as Beijing could snoop on their personal data

  • By Hsu Tzu-ling and Esme Yeh / Staff reporter, with staff writer
  • • 
    • 
    • 

  Information security experts yesterday expressed concern over the risk of personal data being leaked to China, after an app from a Chinese company that uses generative artificial intelligence (AI) went viral on social media in Taiwan.

  After a user provides access to their social media accounts on services such as Instagram, X or LinkedIn, the AI-powered app Roast Master analyzes the user and generates sardonic comments about their MBTI personality type — referring to the personality assessment tool Myers-Briggs Type Indicator popular among younger people — weaknesses and strengths, and fortune predictions for the year.

  Roast Master is offered as one of the services provided by the AI assistant Monica, but it was developed by a Chinese company of the same name, whose Chinese parent company is Butterfly Effect Pte.

  

  The letters AI are pictured on a motherboard in an arranged photograph.

  Photo: Reuters

  As a highlighted feature of Monica, Roast Master is touted as “the ultimate AI roast generator for Instagram, Facebook, Twitter [X]” and other social media despite Instagram and X being banned in China, indicating that the company is aimed at overseas markets.

  Monica’s intention to enter the Taiwan market is also reflected in its numerous advertisements in Taiwan and the app’s language options, which include Traditional Chinese for Taiwan and Hong Kong.

  While using the Web site as a stand-alone service would be less risky, as it reveals less information to the service provider, downloading and registering on the company’s app would allow it to track user behavior, preferences and even access private information, National Taiwan University of Science and Technology Taiwan Information Security Center director said Cha Shi-cho (查士朝) said.

  The Chinese-developed AI assistant could also provide users with misleading information to manipulate their thinking, he said, adding that “it is fine to have fun, but the generated results should not be taken seriously.”

  Democratic Progressive Party Legislator Puma Shen (沈伯洋) said that it is known that these apps could access personal data of their users such as contacts, location and social media accounts.

  There are security risks when it comes to using services provided by Chinese companies or those with Chinese investments, as Chinese law enables the authorities to access their personal data, he said.

  Deputy Minister of Digital Affairs Lin Yi-jing (林宜敬) said the ministry would step up efforts to raise the public’s awareness about the importance of information security.

  Taiwan is a democracy with strong freedoms so the government should not impose strict regulations like authoritarian countries, he said.

  Although AI-powered services present challenges to information security and freedom of speech, the government would only ban a product if it is proven to be a clear security concern, such as by recording voices or activating cameras without user consent, Lin said.

  Cha said people should choose trustworthy companies when using online services and beware of security risks such as data leaks with unidentified service providers.

  Some apps ask users to grant permissions that are not necessary for their functions, including access to microphones and cameras, so users should be aware they should not upload too much information which could reveal their identity, he added.

  
• Most Popular

  • 1

    Sixers down Rockets in overtime

  • 2

    Attack on military convoy in Yemen kills five people

  • 3

    Harris leads Pistons past Celtics in thriller

  • 4

    Alex Honnold is climbing Taipei 101 with no ropes, live on TV. It’s drawing excitement — and concern

  • 5

    Preclearance entry to Japan to start next week