Who has never dreamed of being able to communicate with their cat or dog? After all, our pets being our best friends, wouldn’t it be fascinating to be able to ask them questions and find out what is really going on in their heads? Thanks to advances in artificial intelligence, these imaginings, which back in the day were more the stuff of fantasy films, may be more accessible than it would appear.

A future not so far away

Researchers have been exploring the means of communicating with other animal species for years, but it is since 2023 that the subject has been gaining a lot of media attention. The biologist Michelle Fournet has managed, thanks to AI, to translate a sound emitted by humpback whales. Her research was backed by the Earth Species Project (ESP). 

The non-profit organisation created in 2017 brings together technologists and engineers who work on developing automatic learning systems capable of decoding animal communication. Their goal is to establish bidirectional communication between humans and other animal species, in their own language, by means of computers piloted by AIs, which would act as intermediaries between the two. Katie Zacarian, the CEO and co-founder of the organisation has moreover affirmed, in an article for the World Economic Forum : ‘With this progress, we anticipate that we are moving rapidly toward a world in which two-way communication with another species is likely.’ 

Since 2022, the organisation has been gathering a database to supply AI with. It has already made significant progress in AI interpretation of whale song. This database should, in a few years, enable the interpretation of the languages of primates, birds, mammals, bats and even amphibians, and translate them into human language.

The complexity of animal behaviour as an obstacle

According to the ESP, nature documentaries will no longer need a narrative voice to describe the scenes, because the ‘conversations’ of the animals will be subtitled on the screen. But there is still work to be done before we get that far. And this work is less involved with the technology of AIs than with the immense quantities of data which need to be collected to understand the communication patterns between animals.

The source of this difficulty is the fact that, whilst humans use a complex language based on speech, animals communicate by means of their behaviour, in using a system of signals. A cat, for example, communicates with its peers via the movements of its tail, the position of its ears, the size of its pupils or by a series of sounds. In this case and in many others, simply analysing the sounds in the algorithm generated by the AI would lead to a largely incomplete interpretation of the species’ language. 

To complete the research, it is environmental ecology which is proving to be a major asset. It follows the models and the laws of animal behaviour and thus allows researchers to link movements and behaviours to words and thus to translate precisely the language associated with them.

Whales, the most involved

And in this domain it is the cetaceans who offer the best gateway. ‘Cetaceans are particularly interesting because of their long history—34 million years as a socially learning, cultural species,’ Katie Zacarian, CEO of the ESP explained to the World Economic Forum. ‘And – as light does not propagate well underwater – more of their communication is forced through the acoustic channel.’

The CETI project, a working partnership between researchers specialised in marine biology and artificial intelligence experts, has the particular goal of deciphering the language of sperm whales. The researchers use the sounds recorded over the last fifteen years to decrypt the communications established between these whales. Sperm whales communicate using codas, clicks following different rhythms, which vaguely resemble morse code. 

But decoding the sounds is not enough. To understand the motives behind the communications between the animals, it is vital to understand the social context in which they are exchanged. That involves getting to grips with the ‘where’ and the ‘when’ in addition to the animal in question. 

The advances enabled by the CETI project will only be able to be applied to sperm whales and not to other marine mammals, but the project’s guiding principles will remain relevant. It will all the same be necessary to develop AI translation models specific to each species studied. ‘Where it will be very useful to researchers working on the communication of other marine mammals is finding out where the CETI project has succeeded and where it has failed, to avoid repeating these errors,’ notes the bio-acoustician Michelle Fournet.

AI as an ally for the protection of nature

Just to make sure there is no misunderstanding here, this research will not enable us to have conversations with the species. Its goal is to listen and understand the needs of animals in order to adapt our behaviour. ‘For example, if we know that one sound is essential to enable a whale to feed itself, we can ensure that the ocean is silent at that precise moment so that it can find food. But we need to know which sound is involved,’ explains Michelle Fournet.

In the opinion of the ESP scientists, the breakthroughs in terms of communication with animals, behavioural ecology and automatic learning could impact heavily on conservation efforts. They imagine a world in which we could warn the whales and the dolphins of dangers by sending them messages in their language. In very thoroughly studying the culture of elephants, we could also create more effective conservation corridors which would be communicated to the troops. 

Along similar lines, the AIs will be able to provide environmental organisations with support in monitoring the health of ecosystems , due to the analysis of bio-acoustic recordings. For example, in comparing the sounds of a coral reef in good health to those of a degraded reef, or in measuring the recovery of a tropical forest after a landslide or a fire. 

It is difficult to estimate all the implications which artificial intelligence might have in store for the conservation of the environment, but, placed in the right hands, it offers the very hopeful promise of rebalancing our relationship with nature.  

The other side of the coin

Some scientists nevertheless raise concerns over the dangers these new technologies may represent to the animal species studied. Whilst these tools are currently only tested on species raised in the laboratory, the researcher Olivier Alexandre wonders about the future: ‘their starting point consists of thinking that communicating with animals will be beneficial for the living world. But are we certain of that? If tomorrow industrial applications are found for these technologies – and I have no doubt that they will be found – what will that mean? For the living world, that all the same raises an awful lot of questions.’ 

Concerns all the more legitimate in that the artificial intelligence tools the Earth Species Project is working on are developed in open source. What is there to prevent malevolent organisations appropriating them and using them to their advantage? For example, attracting whales to an area where they are waiting to hunt them? 

Call for projects

A story, projects or an idea to share?

Suggest your content on kingkong.

send your ideas