How AI is decoding the calls of the wild

Hello Nature readers,
Today we learn that an experimental journal is paying peer reviewers in cryptocurrency, get a glimpse of a Milky Way-like galaxy in its youth and investigate efforts to use artificial intelligence to decode the ‘language’ of animals.

Radical journal pays up for peer review

The experimental publication ResearchHub Journal is paying peer reviewers the equivalent of US$150 per review in a specially developed cryptocurrency. The journal is hosted on ResearchHub, a platform backed by crypto entrepreneur Brian Armstrong, which aims to make science more open and efficient. “Getting paid to review is justice,” says molecular-biology consultant Pedro Paulo Gattai Gomes. But some other experts are cautious that the journal might struggle to gain a foothold in research publishing because of its unusual approach, and that the use of cryptocurrency could dissuade cautious academics.

On the trail of a whale’s tail

Using tail-recognition software, researchers have tracked a humpback whale (Megaptera novaeangliae) over 13,000 kilometres — the longest migration ever seen for the species. The whale’s odyssey, from the Atlantic off Colombia to the coast of Tanzania in the Indian Ocean over the course of almost a decade, is unusual for humpbacks, who normally stick to one section of the ocean. “This could be a simple story of a deeply confused whale,” says marine biologist Alexander Werth. “But it’s more likely that this intrepid explorer is a lonely male desperately seeking mates.”

Galaxy could mirror a young Milky Way

Images of a young galaxy captured by NASA’s James Webb Space Telescope have given astronomers a glimpse of what the Milky Way might have looked like when it first formed. The images capture the galaxy, dubbed Firefly Sparkle after its resemblance to the bioluminescent insects, in the process of being assembled from groups of stars around 600 million years after the beginning of the universe.

AI might decode the calls of the wild

Artificial intelligence (AI) is helping scientists to reveal what animals are saying to each other. Researchers have already leveraged the tool to discover that both African savannah elephants (Loxodonta africana) and common marmoset monkeys (Callithrix jacchus) have names. But it’s not as simple as building Google Translate for primate chatter. Such systems need huge amounts of well-defined training data, and it’s still an open question whether animals even have ‘language’ (and what ‘language’ is). Instead, AI can do things such as help researchers comb through recordings, separate sounds and assign them to individual animals. The ultimate goal, for many, is to help protect the creatures they’re studying. “If it were possible for humans to hear from other animals in their own words, ‘Hey, stop fucking killing us’, maybe people would actually do that,” says behavioural ecologist Mickey Pardo.

Satellites alone can’t track methane

Satellites can be effective environmental monitoring tools, but they’re no silver bullet, warns satellite expert Lorna Finman. To accurately measure releases of the potent greenhouse gas methane, satellites should be paired with boots on the ground to verify findings and overcome issues such as weather patterns that can distort readings. “Methane monitoring is too important to leave to one tool alone,” says Finman. “Let’s make sure we get this right.”

‘Make AI for Africa in Africa’

“We cannot expect those who do not understand how we live to build tools for us,” writes artificial-intelligence researcher Nyalleng Moorosi. She argues that making AI systems that are useful for research in Africa is not simply a case of adding more data to Western-built models, which are rarely trained on African languages or culture-specific data. Instead, the research community in Africa deserves opportunities to develop its own AI systems and regulations.