Industry Thought Leadership

Generative AI applications have grown explosively over the years. According to Gartner, generative AI will account for 10% of all data generated and 30% of outbound marketing messages by 2025. Generative AI has been widely used to create text, code, images, videos, and others across industries such as gaming, finance, media & entertainment, biology & chemistry, and others.

In the domain of Text Generation, our Al solutions enable industries such as government, finance and others to increase efficiency through automated customer service, Q&A, and chatbots. In the government sector, our Pangu Government Model helps to build digital government 2.0 with E2E intelligence in government Q&A bot and government copywriting. In the finance industry, our Pangu Finance Model is pre-trained on massive amounts of common-sense knowledge in finance to provide multidimensional content generation and checks such as financial policy document-based Q&A.

In the context of Code Generation, Huawei Cloud combines CodeArts with Pangu to build CodeArts Snap - an intelligent programming assistant for developers to assist in code generation across industries like the financial industry and other sectors. Trained with 76 billion lines of quality code and 13 million technical documents, CodeArts Snap makes generation, Q&A, and collaboration smarter. It generates code through one dialog, automatically comments out and generates test cases in one click, and deploys services/apps with one instruction.

In the area of Image Generation, Huawei Cloud Pangu Models assisted Meitu X-Design in their clothing design through customized exclusive models to produce natural and appropriate clothing designs with the option to switch different scenes. In addition, Huawei Cloud integrated its Pangu model into the MetaStudio digital content production line to create the Pangu Digital Human model that has been pre-trained using 200,000 hours of audio and video data. With this, it takes only 5 seconds to generate a 3D digital human model from a photo, with support for personalized facial adjustments. The "digital human video production" service allows users to create live broadcasts using text, voice, or video input.

In the Scientific Computing field, the Huawei Cloud Pangu Drug Molecule Model can generate millions of new molecular structures and analyze their drug properties to screen potential molecular structures as drugs. This can reduce the costs of trial and error, accelerating the discovery of lead compounds from several years to just one month.

The unprecedented generative AI opportunity
Artificial Intelligence (AI) has become an integral part of today's world, transforming how we live, work, and interact. AI is powering innovations across various industries, shifting from traditional AI models focusing on classification and prediction to AI-Generated Content (AIGC) or Generative AI. According to Gartner, AIGC will account for 10% of all data generated by 2025. Additionally, 30% of outbound marketing messages will be generated by AI by 2025, a significant increase from less than 2% in 2022. We can also expect a major blockbuster with 90% of the film generated by AI in 2025, from 0% of such in 2022. I firmly believe that using AI in various industries will be the next big tipping point, ushering in a new era of ubiquitous connectivity and intelligence and driving innovation and productivity across sectors.

Huawei Cloud Pangu streamlines AI adoption for industry
Implementing AI into business operations requires overcoming a few challenges, such as difficult data acquisition, difficult industry knowledge distillation, and difficult knowledge computing that enterprises need to address. With conventional AI model development, a considerable amount of manpower, customization, and sample training are required for every specific scenario and often take a long time to develop.

Huawei Cloud Pangu Models are trained with hundreds of billions of parameters, outperforming peer models in terms of understanding and generation. The decoupled, hierarchical architecture allows the Pangu models to be quickly adapted to a wide range of downstream tasks. Customers can load independent datasets to train their own models. They can also choose to upgrade foundation models or just upgrade capability sets.

To meet enterprises' data security requirements for AI models, Huawei Cloud provides multiple deployment modes, including public cloud, dedicated zone (public cloud), and hybrid cloud. In this way, enterprises can enjoy the continuous supply of compute resources from the public cloud while keeping data within their own organizations to build their own AI models.

Our models are industry-tailored with specific knowledge and experiences, such as our own business experience over the past 30 years, know-how in more than ten industries, and those from our customers and partners. In addition to an ocean of general knowledge, Pangu models have been pre-trained with open datasets from more than ten industries, including finance, government, meteorology, healthcare, Internet, education, automotive, and retail. The data volume of each industry exceeds 50 billion tokens.

Navigating the policy and ethical landscape of Generative AI
Generative AI is a rapidly developing technology with the potential to impact society significantly, both positively and negatively. As a result, governments and international organizations are developing policies to ensure that Gen AI is used safely and ethically.

For example, the European Union's Artificial Intelligence Act, which is currently being drafted, is expected to be one of the most comprehensive AI regulations in the world. The act will set out requirements for companies that develop, deploy, and use AI systems, including transparency, accountability, and risk assessment.

The regulatory landscape for AI is continually evolving and the level of awareness and compliance varies from one company to another. Therefore, organizations need to stay updated on regional and global policies surrounding AI, as compliance can be complex and rapidly changing.

In addition to being aware of regional and global policies, companies must also be mindful of the ethical implications of using Gen AI. This includes issues such as bias, transparency, and accountability. Companies need to ensure that their AI systems are fair, unbiased, and transparent about how they work. They also need to be accountable for the decisions made by their AI systems.

By being aware of the regional and global policies and ethical implications surrounding the use of Gen AI, companies can help to ensure that this technology is used safely and responsibly.

Generative AI will completely transform creative industries
A Generative AI model is a machine learning-based technology that learns from large amounts of data and generates new content. Compared with traditional AI models, this technology is more flexible and creative and can produce more diverse results. It can be used in natural language processing, image generation, music creation and other fields, bringing new experiences and possibilities to people.

In the field of natural language processing, Generative AI models can be used for text generation, dialogue systems and other tasks. By learning large amounts of text data, generative AI models can generate new text similar to human language and even produce conversations. This provides powerful support for applications such as automated text generation and intelligent customer service. The Pangu NLP has made a significant breakthrough as it was distilled with a large amount of general knowledge in the pre-training phase, allowing the model to embed industry knowledge bases and databases easily to acquire industry know-how efficiently. Huawei Cloud worked with partners to develop the Pangu NLP model for the Arabic language that supports hundreds of billions of parameters with semantic understanding accuracy reaching 95%, becoming No.1 in Arabic language understanding.

In the financial field, companies can use generative AI to create financial models and forecasts to aid in decision-making and risk management. Huawei Cloud provides resilient infrastructure, application modernization technologies that make financial applications agile, and innovative AI and virtual human technologies that build intelligence into businesses. For conventional financial institutions, Huawei Cloud focuses on AICC, digital interaction, and digital banking for them to go digital.

In the field of image generation, generative AI models can be used for image synthesis, image inpainting and other tasks. By learning a large amount of image data, the generative AI model can generate new images similar to real images and even repair damaged photos. This brings new possibilities for artistic creation, image processing and other fields.

In the field of music creation, Generative AI model can be used for music generation, music recommendation and other tasks. By learning a large amount of music data, the generative AI model can generate new music similar to human music, and even personalized recommendations can be made according to users' preferences. This has brought new opportunities for music creation and music promotion.

The essential Generative AI checklist 
A centralized data strategy is essential for adopting Gen AI because Gen AI models need to be trained on large amounts of data. Collecting and preparing for training can be difficult and time-consuming if the data is scattered across different departments and systems. A centralized data strategy makes it easier to manage and access the data, which can help to accelerate the Gen AI adoption process.

Before adopting Gen AI or any AI technology, organizations should consider several key factors to ensure a successful and responsible adoption of AI. These factors include:

• Business Objectives: Clearly Define the business goals and objectives that Gen AI is intended to address, say in improving customer experiences.
• Data Strategy: Gen AI models are trained on data; to ensure data quality and availability, it is essential to have a centralized data strategy in place to ensure that the data is high-quality, well-organized, and accessible to the Gen AI team.
• Ethical considerations: Gen AI raises several ethical concerns, such as bias, transparency, and accountability. Companies must have a plan to address these ethical concerns before adopting Gen AI. We must develop and adhere to a responsible AI framework addressing ethical concerns.
• Regulatory Compliance: It is critical to understand the regional and global regulations that pertain to the use of the AI, especially in sensitive areas such as healthcare or finance. Ensure that the AI system complies with data protection and ethical guidelines.
• Skills & Talent: Gen AI is a complex technology requiring skilled personnel to develop, deploy, and manage. Companies need to have the necessary talent in place before adopting Gen AI.
• Compute Resources: Gen AI models can be computationally expensive to train and deploy. Companies must ensure they have the necessary compute resources before adopting Gen AI.

Generative AI eases talent management challenges
Companies can develop a structured way to experiment with Gen AI to predict the future:

Use Gen AI to forecast demand for skills and talent. Gen AI can be used to analyze data from various sources, such as job postings, social media, and industry trends, to forecast future demand for skills and talent. You can train Gen AI models using historical data to learn patterns, correlations and trends in your workforce. This information can be used to predict future demand for skills and strategic workforce plans that ensure that the company has the right people in the right places with the right skills at the right time.
Use Gen AI to predict employee turnover. Gen AI can be used to predict employee turnover based on various factors, such as employee demographics, performance reviews, and job satisfaction surveys. This information can be used to develop strategies to reduce turnover and retain top talent.
Use Gen AI to simulate different workforce planning scenarios. Gen AI can be used to simulate different workforce planning scenarios, such as the impact of new technologies, changing market conditions, and mergers and acquisitions. This information can be used to develop contingency plans and ensure the company is prepared for any eventuality.