Australia bans Chinese AI platform DeepSeek over security
The Australian government has recently announced a ban on the Chinese-developed artificial intelligence platform, DeepSeek, from all government systems and devices due to national security concerns. This move aligns with similar actions taken by other countries, including the United States and Italy, as these nations also aim to mitigate the risks associated with the platform, especially regarding data privacy and potential foreign interference.
Sarah Sloan, Head of Government Affairs and Public Policy for Palo Alto Networks in Australia, New Zealand, and Indonesia, commented positively on the decision. "Palo Alto Networks welcomes the Australian Government's continued focus on safeguarding Australia's digital ecosystem and ensuring strong protections against evolving cyber threats," she stated. Sloan highlighted the significance of "Direction 001-2025", underscoring the need for vigilance in technology supply chains and collaborations with trusted vendors to enhance cyber resilience.
Concerns primarily stem from DeepSeek's potential exposure of sensitive information to the People's Republic of China (PRC), exacerbated by China's national security laws that compel businesses to provide government access to data. While blocking access to the platform's website and mobile applications seems straightforward, DeepSeek's open-source nature complicates the issue, as it can be executed locally or through cloud environments.
Satnam Narang, Senior Staff Research Engineer at Tenable, elaborated on these complexities. He acknowledged the challenges of completely restricting DeepSeek, given its ability to be run on personal devices and integrated into various apps. Narang also pointed out the absence of robust safety features within the model, which could lead to its misuse for harmful purposes.
In contrast, Melissa Ruzzi, Director of Artificial Intelligence at AppOmni, expressed a more favourable view of another AI platform, Le Chat, developed by European-based Mistral AI. "Given this company is based in Europe, we can expect that they are more diligent overall about the safety of their model compared to models from companies based in China, like DeepSeek," Ruzzi remarked. She noted the additional layer of security provided by the platform's capability to operate in virtual private clouds and its compliance with local data protection regulations.
While DeepSeek has gained attention for outperforming established models like ChatGPT through its innovative training techniques, there are broader implications concerning AI development. Kareem Saleh, Founder and CEO of FairPlay.ai, questions whether size equates to dominance in AI, suggesting that smarter training processes could challenge the assumption that larger models are superior. Saleh also warns of the potential increased computational demands and sustainability challenges resulting from this rapid AI progression.
The ban on DeepSeek underscores the complexities surrounding AI technologies amid growing global concerns over data security, privacy, and the integrity of sensitive governmental operations. As countries grapple with these technological advancements, the emphasis on secure, dependable AI models becomes ever more crucial.
Moreover, this scenario reflects a broader narrative in the technology sector, where the rush for breakthrough innovations is tempered by the necessity for stringent cybersecurity measures. As the international community continues to navigate these challenges, the equilibrium between progression and precaution remains a pivotal focus for governments and technology firms alike.
