A dinosaur park is missing more than a genome
Ancient DNA can reveal extinct life, while gene editing can change a living relative. Recreating selected traits leaves separate questions about the animal and its proposed conservation benefit.

Reading ancient DNA leaves the animal-building problem unsolved. A sequence reconstructed from fragments still needs a living developmental system before there can be an animal to examine.
That gap makes the Jurassic Park question more interesting than a timetable for resurrecting dinosaurs. Several technologies can produce evidence about extinct life or change a living animal. Each achievement needs to be described on its own terms before it can support a claim about restoring a species.
Ancient DNA has a preservation history
A 2022 study reconstructed an approximately two-million-year-old environment in Greenland using environmental DNA. Genetic fragments in sediment helped identify organisms that had lived there. The work recovered information about an ancient community; it did not recover a living cell ready for cloning.
The distinction is especially large for non-avian dinosaurs. The Natural History Museum's explanation describes the lack of usable dinosaur DNA for the familiar resurrection scenario. Those animals disappeared around 66 million years ago, far beyond the time span of the Greenland study.
A single decay figure cannot turn that gap into a precise universal expiry date for DNA. Preservation conditions affect what survives. The original idea that DNA has a fixed half-life everywhere would give readers more certainty than the evidence supports.
Even a well-supported fragment leaves questions about what is missing. Filling a sequence gap with information from a living relative introduces an inference about the extinct animal. As those decisions accumulate, the resulting design needs to distinguish recovered evidence from choices made by its designers.
The living animal supplies the starting point
Colossal's description of its dire-wolf project says it made 20 edits across 14 genes in gray-wolf cells, using information from ancient samples. The company describes the resulting animals as de-extinct dire wolves. The starting cells and selected edits are more specific statements about what the method did.
The peer-reviewed commentary Engineered proxies and the illusion of de-extinction disputes the interpretation of such work as restoring an extinct species. Its argument concerns the difference between selected traits in an engineered organism and the biological identity claimed for it. The company and its critics use different standards for the label, even when discussing the same animals.
A useful description should therefore name the living species used, the traits being altered, and the evidence for the result. A resemblance in coat colour answers a question about appearance. It leaves other questions about development and behaviour to be examined separately.
The same reasoning applies to a proposed mammoth-like animal. Introducing selected traits into a living relative would be an engineering project with its own biological outcome. The extent to which the animal resembles an extinct population would have to be established through observations, not supplied by its name.
A conservation claim needs an ecological test
The IUCN's guiding-principles resource on extinct-species proxies places proxy creation in the context of a possible conservation benefit. That framing asks what the animal would do in an environment and how the proposed benefit would be evaluated.
I would ask a project to define its ecological claim before discussing release. Restoring a lost function is a testable proposal only when the function and the receiving environment are specified. An animal that survives in captivity has passed a different test from one that contributes the proposed benefit after release.
The receiving environment has its own history. A location once occupied by an extinct species may now have different land use, different neighbouring populations, or a different climate. Similarity to a historical animal cannot settle whether introducing the engineered animal is appropriate there.
That creates an obligation to compare options. Habitat protection may address the problem a project claims to solve. Supporting an existing threatened population may be another route. The comparison depends on the particular conservation goal; engineering difficulty alone cannot establish which project produces the greater benefit.
Birth starts another period of responsibility
A live birth would supply evidence about reproduction under the conditions used. The animals would still need long-term care, with observation of health and behaviour throughout their lives. A public announcement made near birth cannot contain those later outcomes.
My concern with a promised park is that the spectacle creates a finish line too early. The scientific questions continue after visitors can see an animal. So do the costs of looking after it when it becomes unsuitable for breeding or when a proposed ecological role fails to materialise.
The people funding the work need to account for that period. Who remains responsible for the animals if the company changes direction? What observations will be released so other researchers can examine the claims? Those are specific ownership questions, not evidence that all such projects must succeed or fail.
The useful achievements will be the ones whose descriptions remain accurate when their names are removed. Recovering ancient sequence information is an achievement. Demonstrating a particular trait in an edited animal is another. A conservation benefit needs its own evidence, collected over the period in which that benefit is supposed to occur.