On Wednesday, Facebook-owned Meta (META.O) released an artificial intelligence model that can detect specific items within a picture, as well as the world’s biggest collection of image comments.
In a blog entry, the company’s research section stated that its Segment Anything Model, or SAM, could recognise things in pictures and videos even when it had not met those items during training.
Objects can be chosen using SAM by tapping on them or typing text instructions. Writing the term “cat” caused the tool to create circles around each of several animals in a picture in one example.
Since Microsoft-backed (MSFT.O) OpenAI’s ChatGPT robot became a phenomenon in the fall, sparking a surge of investments and a scramble to control the sector, big tech firms have been trumpeting their artificial intelligence triumphs.
Although Meta has not yet launched a product, it has revealed several features that use the type of generative AI popularised by ChatGPT, which generates entirely new content rather than simply finding or categorising data like other AI.
According to CEO Mark Zuckerberg, integrating such dynamic AI “creative aids” into Meta’s applications is a goal this year.
Internally, Meta already employs SAM-like technology for tasks such as labelling pictures, regulating banned material, and deciding which posts to suggest to Facebook and Instagram users.
According to the business, the introduction of SAM will increase access to that sort of technology.
The SAM model and information will be made accessible for non-commercial use. Users who submit their own pictures to an associated sample must also promise to only use it for study reasons.
