Bell, Cohere deploy Canadian cybersecurity AI for threat investigation

Bell Canada and AI company Cohere announced the launch of a cybersecurity AI model that will exist within Bell Cyber’s security operations.

The large language model (LLM), announced at Bell’s Cybersecurity Summit in Toronto, was built using Cohere’s enterprise AI technology and is designed to speed up threat investigations and equip analysts with more relevant, consistent analysis. It’s a domain-specific model, which means that it’s trained for a specific industry or function using data from that industry to improve accuracy and performance compared to general-purpose models.

According to the companies, the model was trained for cybersecurity use with data and operational knowledge from Bell Cyber. The press release also mentions that it is currently being integrated into Bell Cyber’s Autonomous Security Operations Centre (ASOC).

Bell says the model is designed to analyze and summarize complex security information, surface the context that is most relevant to an investigation, and support standardized investigative workflows. The company also noted that the model will reduce the time analysts spend gathering and interpreting information, allowing them to focus on validating threats, determining their impact, and directing the response.

Now, it was also noted that Bell Cyber’s ASOC already uses AI and automation to perform tasks like detecting, prioritizing, and containing threats. However, the new model will extend the current capability by “applying enterprise AI grounded in Bell Cyber’s domain knowledge to the language, evidence, and workflows used in investigations, while keeping analysts in control of consequential decisions.”

This partnership was first announced last year alongside a separate Cohere partnership with the Canadian Government. Interestingly, Cohere recently merged with German company Aleph Alpha.