UBTECH Teams Up with HKU to Advance AI Education across the Greater Bay Area
HONG KONG, June 6, 2025 /PRNewswire/ — On June 6, UBTECH Education and the Centre for Information Technology in Education (CITE), part of the Faculty of Education at the University of Hong Kong (HKU), hosted the official launch of the Guangdong-Hong Kong-Macao Greater Bay Area (GBA) Artificial Intelligence Education Development Initiative at HKU. Held under the theme 'AI Empowers Future Education, Technology Drives Innovation in the Greater Bay Area,' the event highlighted the region's commitment to integrating AI into next-generation educational systems. The inauguration of Artificial Intelligence Education & Teacher Development Center was also held in conjunction with the event.
Through collaboration with CITE, UBTECH Education is working to build a pipeline of AI-competent educators. The joint initiative focuses on cultivating AI fluency among teachers and supporting talent development in STEM and innovation through both local and global professional development programs for GBA-based educators.
Building an AI Education Infrastructure in the GBALaunch of the Greater Bay Area AI School Alliance
In recent years, the Hong Kong Special Administrative Region (HKSAR) Government has prioritized artificial intelligence in its development roadmap. The Hong Kong Education Bureau has introduced a dedicated AI Curriculum Module in middle schools, mandating 10 to 14 hours of AI education for students in Secondary 1 through 3 within the ICT curriculum. As part of this broader effort, the HKSAR Government's AI Education Initiative targets reaching 95% of the region's schools by 2025. To date, 82% of primary and secondary schools in Hong Kong have already integrated AI into their teaching programs.
In line with Hong Kong's educational policies, the UBTECH Education-CITE partnership is establishing a collaborative academic-industry platform for AI teaching content and educator training. Plans include building AI demonstration labs in Hong Kong's primary and secondary schools, with further expansion across Guangdong, Hong Kong, Macao—and ultimately into international markets—positioning Hong Kong as a global reference point in AI education. Both organizations will work together to establish AI education and research centers across Hong Kong, with the broader goal of creating a global AI talent certification network that spans more than 100 countries and regions, covering both K-12 and vocational learning pathways. This initiative is designed to support educator professional growth and drive improvements in AI education quality throughout the GBA.
Embodied AI as a Catalyst for STEM and Innovation Learning Debuts in Hong Kong
Tien Kung—the world's first humanoid robot to complete a half-marathon—was showcased at the event. Serving as a powerful symbol of embodied AI, humanoid robots are reshaping the future of education by enabling new forms of experiential and research-based learning. In partnership with the Beijing Humanoid Robot Innovation Center, UBTECH Education is advancing the deployment of embodied intelligence technologies for educational and research applications through an integrated suite of solutions. Anchored by the Tien Kung humanoid robot platform, UBTECH has rolled out the 'Scientific Research and Co-Creation Program,' already adopted by Fudan University, Shanghai Jiao Tong University, Tianjin University and other top-level institutions, with over 100 units ordered. Tien Kung's debut at HKU marks a significant step toward broader adoption across Hong Kong's universities.
At the K-12 level, UBTECH Education is applying its humanoid robotics expertise to enhance public STEM education and innovation capabilities. These efforts are designed to accelerate AI curriculum integration and link scientific instruction with real-world applications. The partnership with CITE also marks the official launch of UBTECH's new instructional model, 'Embodied Intelligence Empowering Science Education and Innovation,' within Hong Kong's education system.
During the roundtable forum at the event, participants from HKU, industry leaders, and educators from schools across the GBA, engaged in a strategic dialogue on the 'Development and Internationalization of AI Education in the Greater Bay Area.'
The alliance between UBTECH Education and the CITE represents the first dual track education-technology collaboration designed to build a robust ecosystem for AI education across Guangdong, Hong Kong, and Macao. The joint effort aims to position Hong Kong as a global leader in AI curriculum development and talent export. Leveraging the GBA as a strategic launchpad, the program seeks to build a transnational AI education and innovation network aligned with the Belt and Road Initiative and broader international efforts. Additionally, the program will support cross-border talent mobility, enhance workforce readiness in AI-related fields across the GBA, and contribute meaningfully to the advancement of Hong Kong's broader innovation and technology agenda.
CONTACT: Hua He, hua.he2@ubtrobot.com
Hashtags
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
The Star
9 hours ago
- The Star
Ispace aborts Moon mission
Space setback: Hakamada (centre) waiting with members of his team for news of the expected landing on the Moon by the company's Resilience craft, in Tokyo. — AFP The country's hopes of achieving its first soft touchdown on the Moon by a private company were dashed when the mission was aborted after an assumed crash-landing, the startup said. Tokyo-based ispace had hoped to make history as only the third private firm – and the first outside the United States – to achieve a controlled arrival on the lunar surface. But 'based on the currently available data ... it is currently assumed that the lander likely performed a hard landing', the startup said yesterday. 'It is unlikely that communication with the lander will be restored' so 'it has been decided to conclude the mission', ispace said in a statement. The failure comes two years after a prior mission ended in a crash. The company's unmanned Resilience spacecraft began its daunting final descent and 'successfully fired its main engine as planned to begin deceleration', ispace said. Mission control confirmed that the lander's positioning was 'nearly vertical' – but contact was then lost, with the mood on a livestream from mission control turning sombre. Technical problems meant 'the lander was unable to decelerate sufficiently to reach the required speed for the planned lunar landing', ispace said. To date, only five nations have achieved soft lunar landings: the Soviet Union, the United States, China, India and most recently Japan. Now, private companies are joining the race, promising cheaper and more frequent access to space. On board the Resilience lander were several high-profile payloads. They included Tenacious, a Luxembourg-built micro rover; a water electrolyser to split molecules into hydrogen and oxygen; a food production experiment; and a deep-space radiation probe. The rover also carried 'Moonhouse' – a small model home designed by Swedish artist Mikael Genberg. 'I take the fact that the second attempt failed to land seriously,' chief executive officer Takeshi Hakamada told reporters. 'But the most important thing is to use this result' for future missions, he said, describing a 'strong will to move on, although we have to carefully analyse what happened'. Last year, Houston-based Intuitive Machines became the first private enterprise to reach the Moon. Though its uncrewed lander touched down at an awkward angle, it still managed to complete tests and transmit photos. Then in March this year, Firefly Aerospace's Blue Ghost – launched on the same SpaceX rocket as ispace's Resilience – aced its lunar landing attempt. The mood ahead of yesterday's attempt had been celebratory, with a watch party also held by ispace's US branch in Washington. After contact was lost, announcers on an ispace livestream signed off with the message: 'Never quit the lunar quest.' The mission had also aimed to collect two lunar soil samples and sell them to Nasa for US$5,000 (RM21,150). Though the samples would remain on the Moon, the symbolic transaction is meant to strengthen the US stance that commercial activity – though not sovereign claims – should be allowed on celestial bodies. — AFP
The Star
9 hours ago
- The Star
This is the phishing scam that gets an identity theft expert in the US 'really, very angry'
Digital thieves are nothing if not persistent and innovative. They keep finding new ways to try to part you from your money. Phishing – where thieves pose as trusted entities or send legitimate looking emails or messages to trick you into giving them access to your accounts – is a widespread method. And it is constantly evolving. 'We've seen phishing go through the roof,' said Eva Velasquez, the CEO of the Identity Theft Resource Center, a San Diego-based national nonprofit. But knowledge is power. So here are three emerging phishing threats to look out for, according to Internet safety experts. All three threats target key parts of people's digital lives: email attachments that lead to fake login pages, multi-factor authentication trickery and deceptive calendar invites. Spending a few minutes reading these pointers could help you avoid getting your ID or money stolen and save you countless hours of dealing with the fallout. HTML attachments that open fake login pages Imagine a busy professional who is in email action mode. In the past 30 minutes on a Saturday morning, he has filled out emailed liability waivers for his seven children's summer camps, filed an expense report for work, answered a secure portal message from the veterinarian about his sick puppy's prescription, skimmed 182 email subject lines and paid five bills from his email inbox, including a car insurance premium and his beloved cheese-of-the-month club. Amid this flurry of inbound emails, ads, invoices and secure messages, he is working on autopilot: opening messages, skimming, clicking and signing in. What a perfect opportunity. Scammers are taking advantage of user distraction – and their trust – by sending emails with HTM or HTML attachments. When clicked, those open a browser file that looks like secure, familiar login page. These pages might look like secure invoice viewers, file-sharing services like DocuSign or Dropbox, or sign-in pages to platforms including Microsoft 365. 'Once the user enters their credentials, they are sent surreptitiously to the attacker's server,' said Vlad Cristescu, the head of cybersecurity with ZeroBounce, a Florida company that helps businesses lower their rate of bounced marketing emails. Why this method is especially insidious: 'There isn't a clickable link in the email, so standard email security filters (which scan for malicious URLs or attachments like PDFs and ZIPs) may not catch it,' Cristescu added. To prevent this, he added, companies should 'restrict HTML attachments unless essential, and users should treat unfamiliar HTML files the same way they'd treat a suspicious link – don't open it unless you're absolutely sure of the sender.' If you do receive incoming communication with an HTML link or attachment, don't engage, said Velasquez, with the ITRC. 'Don't click on links, people. That's the big, overarching message,' she said. Instead, go to the source: call the phone number on the back of your credit card, visit the bank in person. Multifactor authentication tricks If you are one of the many people who uses multifactor authentication, take note. Multifactor authentication is still very helpful and should be used. But Cristescu flagged one way that scammers are taking this tool – which is designed to make people's online accounts more secure – and using it to slither in. As a refresher, multifactor authentication is an added layer of protection that prevents data thieves from logging into your accounts if they have your username and password. It helps ensure that you're the one who typed in your password when you log in, and not some scammer in the Philippines or Poughkeepsie. To use multifactor authentication, you typically download an app, such as Google Authenticator or Microsoft Authenticator. You register your sensitive online accounts, such as Facebook, bank or email, with that app. Then, every time you log into a registered website, the authenticator app generates a new, random code that you enter after your password as a second layer of verification. With the rise of this protection, a new threat has emerged: Scammers who have your username and password can send log-in requests to your authenticator app. Next, the scammer can pose as an IT expert from your workplace and ask you to approve the log-in request. If you fall for it, then boom – the scammer is in. This technique 'exploits a user's frustration and trust in IT. If you're receiving multiple (authenticator) prompts you didn't initiate, that's not a glitch – it's an attack,' Cristescu said. He recommends pausing, never approving these unexpected requests and flagging the interaction with IT. Velasquez added that if you get an authenticator notification and you didn't just log in yourself, 'That is a huge red flag. Stop and address it. Don't ignore it.' Anytime you interact with IT, be sure you're the one initiating that contact, she added. If someone from IT calls or emails you, disconnect and reach back out using a trusted method, such as the same phone number you always dial. Fake calendar invites A third technique data thieves are using is calendar invites. 'I just get really very angry about this one,' Velasquez said. 'It is super hard to detect.' Here's what to look out for. If you use an online calendar like Google calendar or the native iPhone calendar app, you might receive an invitation to an event you didn't see coming. Sometimes these meetings are legitimate. Sometimes, they are not. Scammers 'are now sending meeting requests with malicious links embedded in the invite or 'join' button. These invitations sync directly into calendars and often go unquestioned,' according to ZeroBounce. Scammers use calendar invites because they have 'built-in credibility – they're not usually scrutinised like emails,' Cristescu said. Look for meeting requests from unknown senders and vague event names like 'Sync' or 'Project Review,' he added. In some jobs or roles, meetings routinely get added to calendars by other people –clients, prospects, coworkers, bosses, peers. 'I have gotten these repeatedly,' said Velasquez, with the ITRC. 'Depending on your lifestyle and your job and how you work, these are going to be particularly challenging. They are real calendar invites. The problem is they have malicious software embedded in them – so when you click on portions of them, 'Click to join,' it's like opening an attachment (or) clicking on a suspicious link. It's the same principle.' Cristescu, with ZeroBounce, shared this tip: 'Treat those just like a phishing email. Disable auto-accept where possible and review every invite manually before clicking anything.' Never stop questioning what lands in your inbox or calendar, Cristescu added. 'Always verify the sender's email address, ensure that any link you click matches the legitimate domain, and look out for subtle red flags like spelling errors or unusual formatting.' A big picture pointer 'All three of these (scams) are so common that it has probably happened to every single person reading the article – at least one of them. That's how ubiquitous these are,' Velasquez said. She shared this broader thought: It's less important to know how to respond to each scenario and more important to pause, be skeptical, double check. It's important to be ever more sceptical, because AI makes it easier and easier for thieves to create convincing ruses, Cristescu and Velasquez both said. AI 'really helps with making these phishing offers look and sound so much more legitimate,' Velasquez said. 'And with the amount of data that is out there from public sources and from data breaches, it's very easy to see what relationships people have.' Where you bank, where you do business – that is all fodder for someone to create a copycat page designed to trick you into logging in. Adopt an 'investigator mindset,' Velasquez said. Use this helpful reminder: the acronym STAR, which stands for Stop. Think. Ask questions or ask for help. Reassess. – The San Diego Union-Tribune/Tribune News Service
New Straits Times
13 hours ago
- New Straits Times
US Supreme Court allows DOGE broad access to Social Security data
WASHINGTON: The US Supreme Court on Friday permitted the Department of Government Efficiency, a key player in President Donald Trump's drive to slash the federal workforce, broad access to personal information on millions of Americans in Social Security Administration data systems while a legal challenge plays out. At the request of the Justice Department, the justices put on hold Maryland-based US District Judge Ellen Hollander's order that had largely blocked DOGE's access to "personally identifiable information" in data such as medical and financial records while litigation proceeds in a lower court. Hollander found that allowing DOGE unfettered access likely would violate a federal privacy law. The court's brief, unsigned order did not provide a rationale for siding with DOGE. The court has a 6-3 conservative majority. Its three liberal justices dissented from the order. Liberal Justice Ketanji Brown Jackson, in a dissent that was joined by fellow liberal Justice Sonia Sotomayor, criticised the court's majority for granting DOGE "unfettered data access" despite the administration's "failure to show any need or any interest in complying with existing privacy safeguards." In a separate order on Friday, the Supreme Court extended its block on judicial orders requiring DOGE to turn over records to a government watchdog group that sought details on the entity established by Trump and Musk. DOGE swept through federal agencies as part of the Republican president's effort, spearheaded by billionaire Elon Musk, to eliminate federal jobs, downsize and reshape the US government and root out what they see as wasteful spending. Musk formally ended his government work on May 30. Two labour unions and an advocacy group sued to stop DOGE from accessing sensitive data at the Social Security Administration, or SSA, including Social Security numbers, bank account data, tax information, earnings history and immigration records. The agency is a major provider of government benefits, sending cheques each month to more than 70 million recipients including retirees and disabled Americans. Democracy Forward, a liberal legal group that represented the plaintiffs, said Friday's order would put millions of Americans' data at risk. "Elon Musk may have left Washington, D.C., but his impact continues to harm millions of people," the group said in a statement. "We will continue to use every legal tool at our disposal to keep unelected bureaucrats from misusing the public's most sensitive data as this case moves forward." In their lawsuit, the plaintiffs argued that the Social Security Administration had been "ransacked" and that DOGE members had been installed without proper vetting or training and demanded access to some of the agency's most sensitive data systems. Hollander in an April 17 ruling found that DOGE had failed to explain why its stated mission required "unprecedented, unfettered access to virtually SSA's entire data systems." "For some 90 years, SSA has been guided by the foundational principle of an expectation of privacy with respect to its records," Hollander wrote. "This case exposes a wide fissure in the foundation." Hollander issued a preliminary injunction that prohibited DOGE staffers and anyone working with them from accessing data containing personal information, with only narrow exceptions. The judge's ruling did allow DOGE affiliates to access data that had been stripped of private information as long as those seeking access had gone through the proper training and passed background checks. Hollander also ordered DOGE affiliates to "disgorge and delete" any personal information already in their possession. The Richmond, Virginia-based 4th US Circuit Court of Appeals in a 9-6 vote declined on April 30 to pause Hollander's block on DOGE's unlimited access to Social Security Administration records. Justice Department lawyers in their Supreme Court filing characterised Hollander's order as judicial overreach. "The district court is forcing the executive branch to stop employees charged with modernising government information systems from accessing the data in those systems because, in the court's judgment, those employees do not 'need' such access," they wrote. The six dissenting judges wrote that the case should have been treated the same as one in which a 4th Circuit panel ruled 2-1 to allow DOGE to access data at the US Treasury and Education Departments and the Office of Personnel Management. In a concurring opinion, seven judges who ruled against DOGE wrote that the case involving Social Security data was "substantially stronger" with "vastly greater stakes," citing "detailed and profoundly sensitive Social Security records," such as family court and school records of children, mental health treatment records and credit card information.
