Who doesn’t know the following scene from action films: In order to get past a surveillance camera, it is manipulated by feeding in a short video loop as a signal. The hero or heroine can pass the camera unnoticed, while the security personnel do not notice.
Lenovo showed the new AI assistant AI Now at the CES 2024 electronics trade fair, which will initially be launched in the People’s Republic of China in the coming months and can do a similar trick. AI Now can temporarily represent users in video conferences, for example in Microsoft Teams.
To do this, users simply enter in the AI Now prompt that they would like to leave the conference briefly, as Lenovo showed us in a demonstration. AI Now then records a three-second video from the user’s current camera feed and plays it seamlessly in a loop.
Video loop only lasts three seconds
The entire process only takes a moment; users can step away from the PC without it being immediately noticeable. Since the video recording only lasts three seconds, you shouldn’t overdo it with your absence – and you shouldn’t make any unusual movements when recording.
In our demo, there was also a note in the top right corner that this was a live image. So it might be difficult to cheat your way out of a meeting. After returning, the live camera feed can be activated again; when switching, a small stutter is visible in the display. However, this is unlikely to be noticeable in the context of a video conference.
Lenovo’s AI assistant runs on the PC itself and partially also works offline. AI Now can be used via chat via a sidebar. Users can, among other things, use the assistant to change system settings and, for example, activate the eye-friendly mode or the performance mode. This should be particularly interesting for those users who are not very familiar with the settings of their PC.
The knowledge database is also interesting: Office files and PDFs can be shared with the AI. Users can then use these as a basis for search queries and, for example, have information extracted and displayed. It is also possible to search for content, for example for pages of a screen presentation, which are then displayed directly.
The files in the knowledge base are processed offline and are not sent to a server. The search for content also works in airplane mode, as Lenovo demonstrated to us. They didn’t want to tell us which LLM Lenovo uses for the AI.
Image creation with the help of partners
AI Now can also generate images, but this usually does not work offline. Lenovo uses partners to create the images, but again they couldn’t be named. According to Lenovo, the partners may change depending on the region in which the service will be offered in the future.
On the new Thinkbook Plus Gen 5 Hybrid, however, image generation can also be carried out on the device itself. The notebook consists of a base unit and a removable screen that can be used independently as an Android tablet. The base unit can continue to be used as a Windows PC in parallel via an external monitor.
If images are requested via AI Now on the Android-powered screen, Stable Diffusion is used, which runs on the screen itself. Lenovo has decided to only run the imaging on the tablet unit directly on the device. Anyone who uses the base unit to create images will be connected to the Internet because Lenovo assumes that the tablet unit will be used more for creative purposes.
Lenovo doesn’t want to compete with Microsoft’s copilot
Lenov does not see AI Now as a competitor to Microsoft’s Copilot – the service is intended to be a supplement. AI Now will be the first to be offered for various Lenovo devices in China. It is currently unknown when the service will come to Europe.