There are not just more cameras - each new model has improved functionality. The more cameras, the more information we will be able to see and record, because both the viewing angles and the coverage area will increase. And the more real-time video we have, the more material we can use to train AI.
In addition, cameras are increasingly capable of shooting at high resolutions (eg 4K), so a video stream will require more storage compared to footage captured at lower resolutions. In turn, increasing the detail of the video will expand the possibilities of extracting valuable information from it, which will positively affect the effectiveness of AI algorithms. Moreover, the new cameras broadcast not only the main but also additional low-bitrate video streams used for low-speed monitoring and AI pattern recognition.