Embracing AI for Good: Insights and practices

Summary

Liu Guiqing, an operating officer from China Telecom, delivered a presentation at the AI for Good Summit in Switzerland, focusing on how his company is leveraging artificial intelligence to drive technological transformation and social benefit. He explained that technological revolutions, including the internet boom, cloud computing adoption, and now AI advancement, have created significant opportunities for the telecommunications sector to innovate and grow. China Telecom has responded by transforming from a traditional operator into a technology-driven enterprise, focusing on four strategic areas: network infrastructure, cloud computing, AI development, and quantum security technologies.

The company has developed the Xirang Integrated Intelligent Computing Service platform, which features a five-layer architecture designed to overcome data silos and provide unified access to computing power for corporate and private clients. This platform integrates infrastructure (IaaS), platform services (PaaS), data services (DaaS), and model services (MaaS), supported by China Telecom’s extensive network of over 6,000 edge data centers and 900 large data centers. The company has created comprehensive AI capabilities including the Xincheng large language model and maintains massive datasets containing over 10 trillion communication tokens and 350 TB of cross-industry information.

Liu emphasized that while AI presents tremendous opportunities, it also poses challenges including the widening intelligence divide, security risks, and uncertainties that must be addressed through four key principles: inclusivity, security, sustainability, and universal benefit. He highlighted specific achievements such as 90% accuracy in deepfake detection, 80% accuracy in scam call identification, and the creation of China’s first zero-carbon data center. The presentation concluded with three collaborative initiatives for the global AI community: establishing standardized AI infrastructure frameworks, advancing AI safety governance, and fostering a collaborative industry ecosystem to ensure AI development serves humanity’s best interests.

Keypoints

**Major Discussion Points:**

– **China Telecom’s AI Infrastructure Development**: Liu Guiqing presented the company’s Xirang Integrated Intelligent Computing Service platform, which features a five-layer architecture (SaaS, PaaS, DaaS, MaaS) designed to overcome data silos and provide unified computing power access to corporate and private clients.

– **Leveraging Existing Telecommunications Infrastructure for AI**: The discussion highlighted how China Telecom is transforming from a traditional telecom operator to a technology-driven enterprise by utilizing their extensive network of 6,000+ edge data centers and 900+ large data centers to support AI services.

– **AI Challenges and the “AI for Good” Framework**: Liu addressed key challenges including the widening intelligence divide, security risks, and uncertainties, proposing solutions through four pillars: inclusivity, security, sustainability, and universal benefit.

– **Practical AI Applications and Solutions**: Specific examples were shared including anti-fraud technologies (90% accuracy in deepfake detection, 80% in scam call identification), zero-carbon data centers, and accessibility features for elderly users like voice navigation for ride-hailing apps.

– **Global Collaboration Proposals**: Three key initiatives were proposed: establishing standardized AI infrastructure frameworks, advancing AI safety governance, and fostering collaborative industry ecosystems to ensure AI benefits humanity globally.

**Overall Purpose:**

The discussion served as a presentation at the AI for Good Summit where Liu Guiqing shared China Telecom’s strategic approach to implementing artificial intelligence responsibly, showcasing their technological capabilities while advocating for global cooperation in AI development and governance.

**Overall Tone:**

The tone was formal, professional, and optimistic throughout. Liu maintained a confident and forward-looking perspective when discussing China Telecom’s achievements and capabilities, while also demonstrating responsibility and concern when addressing AI challenges. The presentation concluded on a collaborative note, emphasizing partnership and shared global responsibility for AI development.

Speakers

– **LJ Rich**: Role/Title: Not specified, Area of expertise: Not specified (appears to be moderating/introducing speakers)

– **Liu Guiqing**: Role/Title: Operating officer from China Telecom, Area of expertise: Telecommunications, AI infrastructure, digital transformation, cloud computing, network technologies

Additional speakers:

None identified beyond those in the speakers names list.

# AI for Good Summit Presentation: China Telecom’s AI Infrastructure and Social Responsibility Framework

## Overview

At the AI for Good Summit in Switzerland, Liu Guiqing, Operating officer from China Telecom, presented the company’s approach to leveraging artificial intelligence for technological transformation while addressing social responsibility. The presentation, moderated by LJ Rich, focused on China Telecom’s AI infrastructure development and framework for ensuring AI benefits society.

## China Telecom’s AI Infrastructure Development

### Xirang Integrated Intelligent Computing Service Platform

Liu Guiqing introduced China Telecom’s flagship initiative, the Xirang platform, designed to address fragmented computing resources. As he explained: “we currently have data silos and the computing powers are dispersed, so it is difficult for people to actually get computing power through these five layers. We will be able to more conveniently provide computing power to corporate and the private clients.”

The platform features a multi-layer architecture, though Liu’s explanation of the technical details was somewhat fragmented in the presentation. The infrastructure leverages China Telecom’s network of over 6,000 edge data centres and 900 large data centres, supported by wide bandwidth and low-latency networks for cloud-network integration.

A key component is the Xinhai Intelligent Data Platform within the Data as a Service (DaaS) layer, which incorporates over 10 trillion tokens of communication and over 350 TB of cross-sector information from 14 industries.

### Transformation from Traditional Telecommunications

Liu emphasized how technological revolutions have created opportunities for telecommunications companies to expand beyond traditional services. China Telecom has transformed into a technology-driven enterprise focusing on network infrastructure, cloud computing, AI development, and quantum security technologies.

“AI as a strategic driver of the next wave of technological revolution and industrial transformation is profoundly changing how we live and work whilst redefining communication network architectures, business products, and operational models,” Liu stated.

## AI for Good Framework

Liu addressed both opportunities and challenges in AI development, noting: “AI has brought a lot of development opportunities. At the same time, we must address its emerging challenges, including the widening intelligence divide, escalating security risks, and also uncertainties.”

China Telecom’s framework is built on four principles:

### Inclusivity

Liu argued that “AI should benefit everyone, like water and electricity, like basic utilities,” emphasizing the need to make AI accessible through distributed cloud infrastructure to narrow the digital divide.

### Security

The company has achieved over 90% accuracy in video deepfake detection and over 80% accuracy in scam call identification. Liu noted: “AI-assisted fraud is a big issue, and AI has made frauds more believable. So we have developed AI technologies to identify AI face swapping and AIGCs.”

### Sustainability

China Telecom created China’s first zero-carbon data centre powered entirely by clean energy, and implemented base stations that have achieved over 500,000 tons of carbon dioxide emission reduction.

### Universal Benefit

This focuses on developing accessible products and services, such as voice navigation systems for elderly users to access ride-hailing services, ensuring AI advancement doesn’t exclude vulnerable populations.

## Global Collaboration Proposals

Liu concluded with three initiatives for international cooperation:

**Establishing Standardised AI Infrastructure Frameworks**: Collaboration with stakeholders to promote inclusive and accessible AI technologies through common standards and protocols.

**Advancing AI Safety Governance**: Developing robust design principles, strategic planning processes, and improved risk testing mechanisms across the AI development lifecycle.

**Fostering Collaborative Industry Ecosystems**: Creating partnerships and an open, inclusive global environment for AI development.

## Key Takeaways

China Telecom’s presentation demonstrated how traditional telecommunications companies can transform into technology-driven enterprises through AI adoption. The Xirang platform represents a practical approach to addressing fragmented computing resources and data silos.

The “AI for Good” framework provides a comprehensive approach to responsible AI development, with specific achievements in fraud detection and environmental sustainability showing practical implementation at scale.

The emphasis on global collaboration reflects recognition that AI development challenges require coordinated international responses, with the three proposed initiatives providing a framework for cooperation.

## Conclusion

Liu Guiqing’s presentation outlined China Telecom’s systematic approach to AI development within a framework designed to benefit society. The company’s experience provides insights for organizations implementing AI technologies while maintaining social responsibility and environmental sustainability. The call for global collaboration positions China Telecom as a stakeholder in worldwide AI governance, recognizing that AI’s benefits and challenges require international coordination.

LJ Rich: Operating officer from China Telecom, please welcome Liu Guiqing

Liu Guiqing: This is a guest Ladies and gentlemen, good morning It’s my pleasure to be here in In beautiful, Switzerland for the AI for Good Summit I am honored to share China Telecom’s insights and practice on AI for Good In order to express myself precisely, I will be doing my presentation in Chinese You know technological Revolution has driven profound changes across economic, social, and industrial sectors while creating significant opportunities for the communication sector First, the widespread adoption of the Internet has transformed human interaction patterns In response, the communication industry has innovated network technologies and architectures Ushering in voice and data services boom Second, the rise of cloud computing has enabled universal access to flexible, inclusive, and secure digital information infrastructure Forward-thinking operators have seized this opportunity by building next-generation digital information infrastructure, cloud plus network Unlocking new avenues for business growth So this is the second business growth avenue Third, AI as a strategic driver of the next wave of technological revolution and industrial transformation is profoundly changing how We live and work while redefining Communication network architectures, business products, and operational models China Telecom has embraced the opportunities of this new era of technological revolution and industrial transformation capitalizing on the trends of digitization, networking, AI advancement, and sustainability We are driving technological innovation across four strategic technology areas network, cloud and cloud network convergence, AI, and quantum slash security By delivering emerging services in cloud, network data, AI security, quantum, and digital platforms we are accelerating our transformation and upgrading from a traditional operator into a technology-driven enterprises By doing that we have met the people’s demand for digital lives I would like to show you the new development phase that we are currently in which is smart cloud network convergence we have We are one of the earliest at delivering cloud computing and developing cloud computing infrastructure with the advent of AI age We are moving towards a smart era On the basis of AI infrastructure, we have built the Xirang Integrated Intelligent Computing Service platform so you can see this structure The Xirang platform has five layers So from us to us, to us to us So we have added one more layer, which is the dust layer So now we move from us to us, to us to us, so you have five layers of coordination So why do we have the five layers? because We currently have data silos and the computing powers are dispersed, so it is difficult for people to actually get computing power through these five layers We will be able to more conveniently provide computing power to corporate and the private clients. We provide them with a one-stop shop solution and the Xirang platform So I will show you the five layers from these slides, so from the SAS platform Because we are very good at building data centers, so now we are moving from traditional data center to AI data center Because we have large data centers and edge data centers And we coordinate the two types of data centers. As a traditional telecom company, we have rich resources For example, we have more than 6,000 edge data centers and 900 large data centers So how do we achieve better coordination of cloud data center and edge data centers? Because our network has very wide bandwidth and low latency and cost is also low So how do we leverage the advantages of cloud and network to achieve an integrated digital platform with a combination of cloud and network? And also when it comes to PaaS, we have a trellis platform That is decoupled from computing resources, trellis platform That is decoupled from computing resources, training, inference frameworks, and development tools To effectively give customers a one-stop shop solution to enable customers to get AI services and computing resources Easily and efficiently And also when it comes to DaaS layer, we have established a Xinhai Intelligent Data Platform And we have large data sets from our proprietary data sets and third-party data sets as well as open source data sets So we have a very large data set For example, we have a data repository that is over 10 trillion tokens of communication Alongside over 350 TB of cross-sector information from 14 industries And when it comes to MaaS, we have developed our Xincheng large model So we have integrated models and also industry-specific models And we also have third-party open source models model. So we have this full model, full scale, and fully proprietary platform. You have video, you have language, and you have meaning. So it covers a full range of models. So for example, you have cloud computers and AI phones. So we have a series of utilization scenarios. You know that AI has brought a lot of development opportunities. At the same time, we must address its emerging challenges, including the widening intelligence divide, escalating security risks, and also uncertainties. So how do we address these challenges? So we need to advance AI for good through four aspects, inclusivity, security, sustainability, and universal benefit. Inclusivity means that our AI should benefit everyone, like water and electricity, like basic utilities. So how do we achieve such inclusivity and narrow the digital gap? So we have distributed cloud. So we have a 2 plus 4 plus 31 plus X intelligent computing cloud infrastructure. Also, our network has low latency and low cost. So many applications such as smart driving would be easily supported by our networks. Also, we have many integrated platforms. So through cloud plus data plus AI models plus application solutions, we have created an ecosystem with over 100 digital platforms and implemented 45,000 industrial application projects. Moving on, security is a top concern of all AI experts in the world, and it’s a key for ensuring AI for good. China Telecom has introduced a large language model in security from data to code to operations and attack slash defense and support. We have many core application scenarios such as threat analysis and judgment and intelligent operations. Because we know that AI-assisted fraud is a big issue, and AI has made frauds more believable. So we have developed AI technologies to identify AI face swapping and AIGCs. So we can achieve over 90% accuracy in video deepfake detection and over 80% accuracy in scam call identification, thus offering new solutions and ideas for global anti-fraud efforts. Thirdly, AI in power screen and sustainable development of industries. So we have two cases. For example, in Qinghai Province, we have created the first zero-carbon big data center powered by 100% of clean energy, and the key technologies are now also being scaled up and implemented. And also, some base stations are seeing over 500,000 tons of carbon dioxide emission reduction because of energy reduction efforts for these base stations. And fourthly, inclusive and shared access to intelligent applications because AI products are difficult for the elderly and children, especially for the elderly. So we have developed products such as language navigation models so that they can speak into the app to get a cab, thus making it easier for the elderly to move about. Artificial intelligence is driving unprecedented transformation across the globe. And it is a global responsibility to promote the development of AI. And China Telecom is committed to promoting the healthy and orderly development of AI towards beneficial, secure, and equitable outcomes that ultimately serve humanity’s best interest. To this end, I would like to propose three initiatives. The first, we shall collaborate to establish a standardized framework for AI infrastructure. By harnessing expertise from all stakeholders, we can jointly chart the revolutionary path of artificial intelligence, advance research on next-generation digital infrastructure standards, and facilitate the development of such infrastructure that can be used for and facilitate the development of such infrastructure that integrates cloud network and AI, ultimately promoting inclusive and accessible AI technologies. Secondly, we shall collectively advance the governance of AI safety through robust, top-level design and strategic planning. We can proactively avoid and mitigate potential risks from the source. In addition, we need to improve the AI risk testing and evaluation mechanisms while elevating both safety thresholds and governance capabilities across the AI development lifecycle to ensure AI intelligence progresses in a secure, well-regulated, and controlled manner. Third, we shall foster a collaborative AI industry ecosystem through deepened technological and industrial partnerships. By cultivating a more open, collaborative, and inclusive global environment for artificial intelligence, we can collectively accelerate the realization of AI for good. As China Telecom, as a committed practitioner of AI for good, seeks to deepen collaboration with global partners, closely track technological and industry trends to bridge the intelligence divide, and relentlessly advance AI for good. Thank you. I would like to thank ITU and our global ecosystem partners for your support and your assistance, and I would like to wish this forum a complete success. Thank you very much.

LJ Rich: That was incredible. Thank you very much to Liu Guiqing for a

Agreement points

Appreciation for comprehensive AI presentation

Speakers

– Liu Guiqing
– LJ Rich

Arguments

China Telecom has built the Xirang Integrated Intelligent Computing Service platform with five layers (SaaS, PaaS, DaaS, MaaS) to provide one-stop computing power solutions

Recognition of Liu Guiqing’s comprehensive presentation on China Telecom’s AI initiatives

Summary

Both speakers participated in a successful presentation exchange where Liu Guiqing delivered comprehensive insights on China Telecom’s AI initiatives and LJ Rich acknowledged the quality and thoroughness of the presentation

Topics

Sociocultural | Development

Similar viewpoints

Both speakers demonstrate alignment on the value and importance of sharing comprehensive AI knowledge and initiatives at international forums like the AI for Good Summit

Speakers

– Liu Guiqing
– LJ Rich

Arguments

China Telecom has built the Xirang Integrated Intelligent Computing Service platform with five layers (SaaS, PaaS, DaaS, MaaS) to provide one-stop computing power solutions

Recognition of Liu Guiqing’s comprehensive presentation on China Telecom’s AI initiatives

Topics

Development | Sociocultural

Unexpected consensus

No unexpected consensus identified

Speakers

Arguments

Explanation

The interaction was primarily a presentation format with one main speaker (Liu Guiqing) and acknowledgment from the moderator (LJ Rich). There were no opposing viewpoints presented that could lead to unexpected consensus.

Topics

Overall assessment

Summary

The discussion showed complete alignment between the presenter and moderator, with Liu Guiqing presenting China Telecom’s comprehensive AI for Good initiatives and LJ Rich providing positive acknowledgment

Consensus level

Very high consensus level due to the presentation format. The implications are limited as this was not a debate or multi-perspective discussion, but rather a knowledge-sharing session that successfully conveyed China Telecom’s AI infrastructure developments and AI for Good principles to an international audience.

Different viewpoints

Unexpected differences

Overall assessment

Summary

No disagreements identified in the transcript. The discussion consists primarily of Liu Guiqing presenting China Telecom’s AI initiatives and LJ Rich providing acknowledgment.

Disagreement level

No disagreement present. This appears to be a presentation format rather than a debate or discussion with opposing viewpoints. LJ Rich’s role was limited to introducing and thanking the speaker, with no substantive arguments or counterpoints presented.

Partial agreements

Similar viewpoints

Both speakers demonstrate alignment on the value and importance of sharing comprehensive AI knowledge and initiatives at international forums like the AI for Good Summit

Speakers

– Liu Guiqing
– LJ Rich

Arguments

China Telecom has built the Xirang Integrated Intelligent Computing Service platform with five layers (SaaS, PaaS, DaaS, MaaS) to provide one-stop computing power solutions

Recognition of Liu Guiqing’s comprehensive presentation on China Telecom’s AI initiatives

Topics

Development | Sociocultural

Key takeaways

China Telecom has successfully transformed from a traditional telecom operator to a technology-driven enterprise by embracing AI, cloud computing, and network convergence

The company has developed a comprehensive five-layer Xirang platform (SaaS, PaaS, DaaS, MaaS) that provides integrated AI computing services and addresses data silos

AI for Good must be implemented through four key principles: inclusivity (making AI accessible like basic utilities), security (protecting against AI-enabled threats), sustainability (reducing environmental impact), and universal benefit (ensuring accessibility for all users including elderly)

China Telecom has achieved significant technical milestones including 90% accuracy in deepfake detection, 80% accuracy in scam call identification, and zero-carbon data centers

Global collaboration is essential for AI development and requires standardized frameworks, robust safety governance, and inclusive industry ecosystems

Resolutions and action items

Establish standardized frameworks for AI infrastructure through collaboration with global stakeholders

Advance AI safety governance through improved risk testing and evaluation mechanisms across the AI development lifecycle

Foster collaborative AI industry ecosystems through deepened technological and industrial partnerships

Continue scaling up zero-carbon data center technologies and energy reduction efforts

Deepen collaboration with global partners to bridge the intelligence divide and advance AI for good

Unresolved issues

Specific implementation details and timelines for the three proposed global collaboration initiatives were not provided

Technical specifications and standards for the proposed AI infrastructure framework remain undefined

Mechanisms for international coordination and governance of AI safety standards were not detailed

Funding models and resource allocation for global AI collaboration efforts were not addressed

Specific metrics and benchmarks for measuring ‘AI for Good’ success globally were not established

Suggested compromises

None identified

AI as a strategic driver of the next wave of technological revolution and industrial transformation is profoundly changing how we live and work while redefining Communication network architectures, business products, and operational models

Speaker

Liu Guiqing

Reason

This comment is insightful because it positions AI not just as another technology tool, but as a fundamental force that is reshaping entire industries and human behavior patterns. It demonstrates a macro-level understanding of AI’s transformative power beyond simple automation or efficiency gains.

Impact

This comment established the foundational framework for the entire presentation, setting up AI as the central theme around which all subsequent technical and strategic discussions would revolve. It shifted the conversation from traditional telecom services to AI-driven transformation.

We currently have data silos and the computing powers are dispersed, so it is difficult for people to actually get computing power through these five layers. We will be able to more conveniently provide computing power to corporate and the private clients.

Speaker

Liu Guiqing

Reason

This comment is thought-provoking because it identifies a critical infrastructure problem – the fragmentation of computing resources – and presents a concrete architectural solution. It reveals the practical challenges behind AI deployment at scale and how traditional telecom companies can leverage their infrastructure advantages.

Impact

This comment transitioned the discussion from high-level AI concepts to specific technical implementation challenges and solutions, introducing the Xirang platform as a practical response to real-world infrastructure limitations.

AI has brought a lot of development opportunities. At the same time, we must address its emerging challenges, including the widening intelligence divide, escalating security risks, and also uncertainties.

Speaker

Liu Guiqing

Reason

This comment is particularly insightful because it acknowledges the dual nature of AI advancement – both its tremendous potential and its inherent risks. The concept of an ‘intelligence divide’ is especially thought-provoking as it parallels the digital divide but suggests even more profound societal implications.

Impact

This comment marked a crucial turning point in the presentation, shifting from technical capabilities and business opportunities to ethical considerations and social responsibility. It introduced the ‘AI for Good’ framework that would structure the remainder of the discussion.

AI should benefit everyone, like water and electricity, like basic utilities… AI-assisted fraud is a big issue, and AI has made frauds more believable. So we have developed AI technologies to identify AI face swapping and AIGCs.

Speaker

Liu Guiqing

Reason

This comment is deeply thought-provoking because it presents AI as a fundamental utility while simultaneously acknowledging that AI itself creates new categories of threats that require AI solutions. The metaphor of AI as a utility like water and electricity suggests universal access and essential infrastructure, while the fraud detection example shows the arms race dynamic in AI development.

Impact

This comment introduced a complex paradox that deepened the discussion’s analytical level – the idea that AI solutions are needed to combat AI-generated problems. It also elevated the conversation to consider AI’s role in society’s basic infrastructure.

We shall collaborate to establish a standardized framework for AI infrastructure… collectively advance the governance of AI safety… foster a collaborative AI industry ecosystem

Speaker

Liu Guiqing

Reason

These three proposals are insightful because they recognize that AI development cannot be effectively managed by individual companies or nations alone. The call for global collaboration on standards, safety governance, and ecosystem development acknowledges the inherently global nature of AI challenges and opportunities.

Impact

These concluding proposals shifted the discussion from China Telecom’s specific initiatives to broader global cooperation needs, positioning the company as a stakeholder in worldwide AI governance rather than just a technology provider.

Overall assessment

While this transcript represents primarily a monologue presentation rather than an interactive discussion, Liu Guiqing’s key comments effectively structured a comprehensive narrative that moved from technological transformation to practical implementation challenges, then to ethical considerations, and finally to global cooperation needs. The most impactful comments were those that revealed the complexity and dual nature of AI development – showing both tremendous opportunities and significant risks that require collaborative solutions. The speaker’s ability to connect technical infrastructure details with broader societal implications created a multi-layered discussion that addressed AI from business, technical, ethical, and policy perspectives. The progression of ideas demonstrated how AI development in telecommunications requires not just technical innovation but also social responsibility and international cooperation.

How can the standardized framework for AI infrastructure be practically implemented across different countries and organizations with varying technological capabilities?

Speaker

Liu Guiqing

Explanation

This is important because Liu proposed establishing a standardized framework but didn’t provide specific implementation details or address how to handle technological disparities between different stakeholders globally.

What specific mechanisms and metrics should be used for AI risk testing and evaluation across the development lifecycle?

Speaker

Liu Guiqing

Explanation

This requires further research as Liu mentioned the need to improve AI risk testing and evaluation mechanisms but didn’t specify what these mechanisms should look like or how they would be standardized globally.

How can the scalability and replication of zero-carbon data center technologies be achieved in different geographical and regulatory environments?

Speaker

Liu Guiqing

Explanation

This is important for sustainability goals as Liu mentioned their zero-carbon data center in Qinghai Province and that technologies are being scaled up, but didn’t address the challenges of implementing this globally.

What are the specific technical approaches and accuracy benchmarks for AI-assisted fraud detection that can be standardized across different telecommunications networks?

Speaker

Liu Guiqing

Explanation

This requires further investigation as Liu mentioned achieving 90% accuracy in video deepfake detection and 80% in scam call identification, but didn’t provide technical details that could be adopted by other operators globally.

How can the coordination between large data centers and edge data centers be optimized to ensure consistent AI service delivery across different network infrastructures?

Speaker

Liu Guiqing

Explanation

This is important for the practical implementation of AI services as Liu described their approach with 6,000 edge data centers and 900 large data centers but didn’t detail how other operators with different infrastructure could achieve similar coordination.