(Original title: GPT-4 enters the field of network security, Microsoft “AI Family Bucket” adds Security Copilot)

News from the Financial Associated Press on March 29 (edited by Shi Zhengcheng)On Tuesday (28th) local time, Microsoft, which occupies the high ground of the AI ​​industry, applied GPT-4 technology to the field of network security and launched a new product called Security Copilot.

(Source: Microsoft official website)

Perhaps it is because there have been a lot of discussions regarding “AI leading to unemployment” in the news media recently, so Microsoft emphasized at the beginning of the announcement that this product does not have the possibility of “robbing people’s jobs” because the gap in skilled network security experts in the world reaches 3.4 million people, these people have long faced “asymmetric and highly sophisticated” attacks by attackers, and urgently need the help of AI assistants.

Microsoft introduced,Security Copilot combines GPT-4 and self-trained AI model. This AI robot can help security personnel quickly collect the latest security incident information on the company’s network, dig out potential threats, and quickly find common vulnerabilities. The Microsoft security team receives 65 trillion pieces of information related to network security threats every day, and employs 8,000 security engineers for analysis and research. At the same time, Microsoft, which is good at using “banknote capabilities”, has also broadened its accumulation in the security field by acquiring companies such as RiskIQ and Miburo.

Similar to most GPT family products,Microsoft’s Security Copilot is also presented in the form of prompt word dialog box. Engineers can ask AI for help by inputting natural language, web links, json files, etc.

(Source: Microsoft)

Sitting on the most powerful AIGC capability on the planet, Microsoft’s Security Copilot naturally also provides visualization tools.Security personnel can let the robot generate demo files with one clickshowing the path of the security threat.

(Source: Microsoft)

Microsoft emphasized that although the company will continue to train AI to improve its capabilities, user data will not be used in areas outside its own enterprise. Microsoft also emphasizes that Copilot has the possibility of making mistakes, so it also provides users with convenient feedback means.

Similar to Office’s Copilot, Microsoft did not announce the pricing and launch time of Security Copilot today, only saying that it will let a small number of users try it first. In the future, this product will also join a series of software in the Microsoft security line. The company disclosed in January this year,Microsoft’s security business will generate a total of $20 billion in revenue in 2022。

Frank Dickson, vice president of security business at IDC, an industry research organization, interpreted that if Security Copilot is bound with Microsoft’s Sentinel or other products, it will have a huge impact on the purchasing decisions of target customers. Obviously, competing products like Palo Alto Networks will definitely launch their own security chatbots in the future, but being the first to release a product means that Microsoft can lead the way.