"Enhancing Access to Advanced Technology without Extra Cost"
On July 31, Zhou Hongyi, founder of 360 Group, announced at the 12th Internet Security Conference (ISC.AI 2024), "I have always stated that we need to demystify large models and provide their standard capabilities for free to users purchasing 360 standard products."
In 2005, 360 entered the internet sector by offering free antivirus software and has since adapted its products and strategies through the eras of PC, mobile internet, and artificial intelligence to address evolving security challenges. Zhou believes that advancements in digitalization and intelligence have made the world more vulnerable. He emphasizes that cybersecurity is shifting from human-to-human conflicts to battles between humans and machines, ultimately evolving into competition among machines. To tackle these challenges, we must use AI to address security concerns that also arise from AI technology.
The ISC.AI 2024, co-hosted by the China Internet Association, Chinese Association for Artificial Intelligence, China Software Industry Association, and the 360 Internet Security Center, marked a significant transition from ISC to ISC.AI. Wu Shizhong, an academician at the Chinese Academy of Engineering, remarked, “The new security challenges brought by AI technologies are already a reality, and research into the security of large models is just beginning. Both security research and products must keep pace with technological advancements and innovative applications.”
Wu Jiangxing, director of the National Digital Switching System Engineering Technology Research Center (NDSC), highlighted the severe imbalance of responsibilities and risks in AI application system security. He proposed that an endogenously secure architecture could identify or correct security issues within AI systems. It is crucial that AI applications incorporate necessary diversity and relatively accurate principles to achieve security outcomes, especially in less trustworthy environments. The intelligence era demands a careful selection of technological pathways to secure AI applications.
From the perspective of a security firm, Zhou Hongyi asserted, "Traditional security approaches are inadequate for addressing AI-related security problems, as AI is inherently more sophisticated than conventional software. To combat the challenges posed by AI, we must apply methods that utilize models to counteract these threats. Our strategy remains rooted in providing these capabilities at no cost.” For users purchasing 360 standard products, the company has decided to offer large model capabilities for free. Zhou stated, “We intend to democratize access to these advanced technologies and drive a transformative productivity shift in the security industry. Our goal is to prevent large models from becoming exclusive tools for a select few companies."
The shift from reduced pricing to free offerings indicates that the competition in the large model space is far from over. From the viewpoint of third-party research institutions, Wang Zhongyuan, president of Zhiyuan Research Institute, noted, “For B-end users choosing large models, effectiveness is the top priority. Ultimately, the aim is to see if large models can address real business needs, as prices will eventually decrease.”
Regarding effectiveness, Zhou cited a major state-owned enterprise with an asset scale of 150,000 servers. Previously, asset inventory took around 30 days, but by integrating large models with their security systems, this process is now completed in just two days. Tasks involving 25 data sources and over 700 security devices, which once took 15 days, can now be achieved in two days. With a team size of three to four hundred, where half were previously engaged in alert analysis, each team member can now handle up to 200 incidents daily, resulting in a 300% increase in efficiency. “This is the outcome we aspire to achieve—tying large models directly to production and business operations,” Zhou concluded.
Ranking
