Huawei denies copying Alibaba’s AI model amid China’s heated competition

Huawei’s AI research division has denied allegations that its Pangu Pro large language model copied elements from Alibaba’s Qwen AI model, following claims made in a technical paper posted on GitHub. The controversy highlights growing tensions in China’s competitive AI landscape, where tech giants are racing to develop cutting-edge models following the success of startup DeepSeek’s low-cost R1 model.

What happened: An entity called HonestAGI published a paper claiming Huawei’s Pangu Pro Moe model showed “extraordinary correlation” with Alibaba’s Qwen 2.5 14B model.

• The paper alleged that Huawei’s model was derived through “upcycling” rather than being trained from scratch, suggesting potential copyright violation and false claims about Huawei’s investment in model training.
• Noah Ark Lab, Huawei’s AI research division, rejected these claims on Saturday, stating the model was “not based on incremental training of other manufacturers’ models” and featured “key innovations in architecture design and technical features.”

Key details: Huawei emphasized that its Pangu Pro model represents a significant technical achievement built entirely on its own hardware.

• The company claims it’s the first large-scale model built entirely on Huawei’s Ascend chips, demonstrating the company’s push for technological independence.
• Huawei’s development team said it “strictly adhered to open-source license requirements for any third-party code used,” though it didn’t specify which open-source models it referenced.
• The company open-sourced its Pangu Pro Moe models on Chinese developer platform GitCode in late June to boost adoption among developers.

The competitive landscape: China’s AI market has intensified dramatically since DeepSeek’s breakthrough earlier this year.

• DeepSeek’s open-source R1 model, released in January, shocked Silicon Valley with its low development costs and sparked fierce competition among Chinese tech giants.
• While Huawei entered the large language model arena early with its original Pangu release in 2021, it has since been perceived as lagging behind rivals.
• Alibaba’s Qwen 2.5-14B was released in May 2024 as part of the company’s small-sized model family designed for PC and smartphone deployment.

Different market focus: The two companies target distinct user segments with their AI models.

• Qwen is more consumer-facing and offers chatbot services similar to ChatGPT, positioning it for broader public use.
• Huawei’s Pangu models tend to be used primarily in government, finance, and manufacturing sectors, reflecting the company’s enterprise-focused strategy.

What remains unclear: Several aspects of the controversy lack transparency.

• Reuters was unable to contact HonestAGI or determine who is behind the entity that made the allegations.
• Alibaba did not immediately respond to requests for comment about the claims involving its Qwen model.
• The technical details of the alleged similarities between the models have not been independently verified by third parties.