Scientists Use AI and Gene Editing in Bid to Bring Back the Dodo, Extinct for 350 Years

A team of scientists is working to bring the extinct dodo back to Earth. The flightless bird disappeared from the planet more than 350 years ago, and it could possibly be seen walking the Earth again in the early 2030s. The work is being carried out by Colossal Biosciences, based in Dallas in the United States.

A report published in Time said very little ancient dodo DNA has survived, and what remains has decayed over the centuries. Artificial intelligence, with its analytical power and ability to work with vast amounts of data, could play an effective role in filling those gaps in the genetic information.

Ben Lamm, chief executive and co-founder of Colossal Biosciences, said in a statement: “Bringing back an extinct animal is less a biology problem than a data problem. AI learns how regulatory DNA determines the structure of living birds, and then identifies the genetic switches that played a role in creating the dodo’s distinctive physical structure.”

Colossal Biosciences says it demonstrated its de-extinction capability in 2025, when it announced it had sequenced the ancient DNA of the long-extinct dire wolf. Using that information, the company edited the genome of the living gray wolf, reviving the Ice Age dire wolf more than 10,000 years after its extinction. Its list also includes the woolly mammoth, the Tasmanian tiger, the moa and the bluebuck.

For the dire wolf, researchers first collected cells from a living gray wolf and made 20 edits across 14 genes, aiming to reproduce features such as the dire wolf’s white coat, larger size, powerful jaw and metabolism suited to cold climates. The edited cell’s nucleus was placed into a wolf egg cell, the resulting embryo was implanted in a large domestic dog, and just over 60 days later the surrogate mother gave birth to a wolf pup.

Bringing back a bird is harder, because a bird embryo develops inside an egg rather than in the mother’s womb. To revive the dodo, researchers will first collect primordial germ cells from the Nicobar pigeon, the dodo’s closest living relative; these immature cells later develop into eggs and sperm. The germ cells will be edited to carry dodo traits and inserted through the shell of a domestic chicken egg into the tiny embryo, where they will migrate to the embryo’s reproductive organs. When the edited female chicken grows up and breeds with an edited male, she will lay eggs from which not chickens but dodos will hatch.

Anna Kitte, director of bird species at Colossal Biosciences, said: “In an ideal situation, if we had hundreds of Nicobar pigeon eggs, we could use the pigeon itself as the surrogate species. But Nicobar pigeons breed in a specific season, lay eggs only a few months a year, and usually lay just two eggs at a time. Chicken eggs, on the other hand, are easy to obtain — some chickens lay an egg every day — so we can get far more surrogate eggs from chickens than from Nicobar pigeons.”

The biggest challenge is the dodo’s beak. More than 650 regulatory elements are linked to beak shape, and around 1,700 elements are linked to beak thickness and width. The researchers are not working only with genes; they are analysing regulatory elements — the parts of the genome that determine when a gene switches on and off. To address the data gap, Colossal has built a new AI platform called Genomorph.

To build Genomorph, the scientists created a genetic map showing how the genetic switches controlling facial development turn on and off as chicken, quail, duck and pigeon embryos grow. They then compared the genomes and skull shapes of more than 70 bird species, including the Nicobar pigeon. The work produced a list of about 2,350 genetic switches that may help explain the dodo’s unusual beak shape, and researchers are now testing those switches one by one.

The work rests on another of the company’s achievements: Colossal has managed to assemble a nearly complete dodo genome from preserved samples. Beth Shapiro, the company’s chief science officer, said: “For an ancient genome, this is an extraordinary sample. It has a few small gaps, but far fewer than are usually seen with ancient DNA. We have actually lost very little from the dodo genome.”

The genetic instructions for all the dodo’s traits are hidden within the roughly two billion letters of its genome, which is why Colossal’s scientists advise patience — bringing the dodo back could take about five years. Shapiro said in a statement that a method developed for the dire wolf is now contributing valuably to research on another species, and that each scientific advance becomes part of a technology platform that can move future projects faster.