How AI Is Helping Us Talk to Whales
Conservation Through Conversation
Back in June, here on AI Business Report, we released an article titled Move Over Dr. Dolittle, Now AI Can Talk to Animals, wherein we discussed the possibility of artificial intelligence being used to communicate across species. Now, we’d like to get a little more specific and discuss how AI is helping researchers talk with whales.
Unlike our previous article on cross-species communication’s broad possibilities, talking with whales is, apparently, a likely eventuality. In this post, we’ll explore the researchers behind AI-whale communication, how AI learns languages, and the potential ethical implications of this endeavor. Keep reading to learn more.
Project CETI and Earth Species Project
The two major researchers behind AI-whale communication are Earth Species Project, which we discussed in our previous article about animal-AI communication, and Project CETI, a research institute dedicated to preserving and understanding the animal kingdom. According to their website, Project CETI’s ultimate goal is to use technology “to amplify the magic of our natural world” and “bring us closer to nature.”
In the 1960s, it was CETI researchers who co-discovered that whales sing to one another, sparking the “Save the Whales” campaign. Earth Species Project, meanwhile, is a nonprofit science-tech company that leverages AI to decode “nonhuman” language. Together, the two companies are now using AI to interpret whale songs.
How AI Learns Languages?
Artificial intelligence models, specifically Large Language Models (LLMs) like ChatGPT and Bard, learn languages by digesting huge amounts of written data. From this information, they’re able to identify which words go together and why, based on the wider context of the information they are given.
As a result, LLMs are not restricted to just English, despite most popular models primarily using it. For example, Chinese models use Mandarin, while models developed in non-English speaking countries can use other languages. So, if an AI developer wanted to build a French-speaking AI model, they’d need to train the model on vast quantities of French language content.
Here, we begin to see how it’s possible for LLMs to understand nonhuman languages. In the case of whales, researchers can train AI models on large quantities of whale songs and noises, enabling the AI algorithm to identify associations and correlations between certain noises to build up a general idea of the species’ language.
Therefore, the challenge is not so much in recording and understanding whale language, but translating it into something humans can understand. While it may be easy to identify when two distinct whale noises are associated with one another, translating this requires a broader understanding of context.
The problem of context is the main issue researchers are still grappling with. Thanks to AI, however, scientists are in an excellent position to continue their work deciphering the whale language, using AI models as a jumping-off point.
Should We Talk With Whales?
There’s a long road ahead before humans are able to talk with whales, yet the question isn’t just of “can,” but of should we? Of course, communicating with whales could lead to several benefits, such as better preservation and easier tracking for scientific purposes. For example, through conversation, we may be able to learn of injured or endangered whales whom scientists can then support.
However, humans are notorious for our abuses of power. What if, for instance, this technology fell into the wrong hands and was used to locate whales for fishing purposes? Well, then we might have an even bigger whale population crisis.
There’s also always the chance that AI gets it wrong, miscommunicating whale language to researchers. After all, AI hallucinations – where AI models simply make information up to satisfy operations – are still a common issue among LLM algorithms. So, miscommunications aren’t unlikely.
Human intervention into wildlife almost always comes with wider impacts that aren’t always fully considered due to the relentless march of innovation. And, with AI regulations slow on the uptake and the technology still in the development phase, maybe we should leave whales alone. Haven’t we already meddled enough with the natural world?
Thanks for reading.
If you enjoyed this article, please subscribe to receive email notifications whenever we post.
AI Business Report is brought to you by Californian development agency, Idea Maker.
Sources: