As AI-powered machines edge closer to everyday life, the growing presence of robotic pet dogs on pavements has flagged a legal gap in accident liability, according to experts.
While applications of robots have begun in the city, ranging from government departments to building security, these “strange newcomers” seem to pose a potential security risk to the public. The incidents range from a robot shocking an elderly woman in Macau to another hurting a Hong Kong tech expert's foot.
A recent scene in Tseung Kwan O has further sparked safety concerns, where a man walking his robot dog nearly triggered a crash as a cyclist braked so hard to stare that he almost caused a collision with another incoming bicycle.
Several have further highlighted concerns over camera-equipped robot dogs that could invade privacy, or even commit upskirting without victims knowing.
Francis Fong Po-kiu, honorary president of the Hong Kong Information Technology Federation, recounted a time that an out-of-control robot stepped on his foot, describing it as “being hit with a hammer.”
Fong believed that Hong Kong needs urgent regulation based on a robot's power, speed, and level of autonomy.
Hard to find the one accountable for: legal experts
Under current law, if an AI-powered robotic dog injures a passerby, the owner may face civil liability—especially if they were controlling the device at the time.
However, barrister Albert Luk Wai-hung warned of legal gaps when a machine acts autonomously, likening it to dangerous driving without a driver, which poses prosecution challenges.
Luk pointed to further challenges to anonymous operators, who could control a robot remotely from another location—or even another country—making investigation nearly impossible.
He calls for early public consultation and dedicated legislation governing robots.
Legal constituency lawmaker Nicholas Chan Hiu-fung said responsibility in a robot-related accident can fall on multiple shoulders—the operator, the programmer, or the manufacturer.
While many robot dogs available online from overseas platforms are required to meet Hong Kong's electrical safety standards, he admitted cross-border legal action remains a major hurdle.
However, Chan urged caution against over-regulation, saying “We must balance safety with the need to advance technology.”
Currently no legislation in HK
In response, the Innovation, Technology and Industry Bureau noted that there is currently no specific legislation in Hong Kong to regulate the application of robotics.
However, it is stated that relevant projects must follow regulations in different sectors, citing examples of robots that are overseen by the Department of Health and Development Bureau.
As for AI-related crimes, the department indicated that existing ordinances could cover privacy and cyber offenses.
Acknowledging the legal gaps, the department stated that the Digital Policy Office has developed the Ethical AI Framework and a generative artificial intelligence technical application guideline for the development of AI technology.
According to reports, the European Union regulated the use of AI technology with the world's first comprehensive AI law in August 2024, assessing and banning AI applications based on a risk-based classification system.
As for Japan, its AI law, which became effective last month, requires the government to promote responsible development and allows non-binding investigations when citizens' rights are harmed.
South Korea enforced its AI law in January 2025, requiring human oversight for AI used in nuclear safety, transport, and healthcare, whereas AI-generated content must be clearly labeled if it could be mistaken for reality.
Meanwhile, China published its first governance guidelines for humanoid robots and related standard systems in February 2026, covering the entire chain of the humanoid robot domain.
𝗗𝗼𝘄𝗻𝗹𝗼𝗮𝗱 𝗧𝗵𝗲 𝗦𝘁𝗮𝗻𝗱𝗮𝗿𝗱 𝗔𝗽𝗽 ↓
