DeepMind AI used to track and conserve vulnerable Serengeti species

Many have warned against the potential threats of AI and others are singing its praises for how it can improve the world and lives of people around us. DeepMind, an AI company owned by Google’s parent company Alphabet, wants very much to be part of the latter.

The company is partnering with conservationists and ecologists on a project that uses artificial intelligence algorithms to detect and count animals in “millions” of photos taken over the past nine years in Tanzania’s Serengeti National Park. DeepMind have developed a machine learning model that can accurately identify most animals found in this part of the world, saying that it will shorten the process by at least nine months.

Read: Sony is replacing drummer with AI

Usually volunteers would need to manually and painstakingly label each photo individually.

Taking into consideration how difficult it can be to capture photos of moving animals without motion blur, the AI will be able to detect these animals even when they appear as a relatively blurry image.

According to DeepMind this technology should soon be usable in the wild as well. It is developing a version of its Serengeti model that would be pre-trained and deployable with little resource, including modest hardware and internet connectivity. They are hoping to validate its models soon by rolling them out in the field to gauge their progress in real conditions.

Most conservationists are of the opinion that this could greatly improve their efforts to monitor and track populations of animals. This makes the conservation effort a lot easier, especially in areas like the Serengeti where growing human activity poses a threat to the ecosystem. While AI can’t be the solution to the problem, it can certainly be a feather in the bow of conservationists.