It was vision, not limbs, that led fish on to land, according to a groundbreaking study of evolution published this week.
But let’s hope it’s neither vision nor limbs that leads artificial intelligence to make its own evolutionary breakthrough.
Boston Dynamics has recently unveiled a robot that puts our human limbs to shame – and now Google has announced a major step forward for AI vision.
With the aid of machine learning, the tech giant’s Cloud Video Intelligence API can recognise objects in videos, and it’s frighteningly astute.
Announcing the breakthrough at Google’s Next Cloud Conference, the company’s chief scientist for cloud AI, Fei-Fei Li, showed how the API could ‘watch’ a commercial and identify not only a daschund in the ad, but also that the video was a commercial, The Verge reported.
In a further example in a blog, Google showed how the API was able to identify that a video (see above) featured a tiger and that it’s living in a zoo.
The tool will let people search through videos for specific objects, Google said. Such a task previously required manual tagging.
The API follows in the footsteps of a number of other Google tools that empower AI. Li explains: “These APIs let customers build the next generation of applications that can see, hear and understand unstructured data.”
It’s impressive and alarming in equal measures.