AIで危ないって言われるバックドアってなんですか？

"Chinese AI is vulnerable to backdoors"—this is a common sentiment, but surprisingly few people can explain what a backdoor actually does. So, what aspects of Chinese AI are truly dangerous, and what aspects are not a cause for concern? The key to the answer lies in the fact that the situation changes completely depending on whether it's used online or locally. Furthermore, the term "backdoor" actually encompasses three completely different things. This video will calmly categorize the dangers of Chinese AI by covering: The real risks of online use What happens when it's run locally The three types of the term "backdoor" We will explain these points using concrete examples such as the National Intelligence Law and sleeper agent research. Please subscribe and like! #ChineseAI #DeepSeek #InformationSecurity #AIRisk #Backdoor