Since last year, a policy shift around animal testing has been carried out simultaneously in multiple countries. The UK government proposed to phase out some animal testing in November 2025, cancel skin irritant animal testing this year, and plan to significantly reduce dog related research by 2030. The long-term vision is to no longer use animals except in "extremely special circumstances". The US Food and Drug Administration (FDA) announced that it will make animal testing an "exception rather than a routine" in drug safety assessments within 3-5 years, and the National Institutes of Health (NIH) stated that it will reduce funding for research that specifically relies on animal models. The EU also plans to release a roadmap for the "de animalization" of chemical safety assessments. The driving force behind this change comes from both ethical and animal welfare demands, as well as the rapid maturity of alternative technologies. According to a recent report on the UK's Nature website, a technology system called "New Methodology" (NAM) is attempting to provide a new "foundation" for medical and toxicological research. The three alternative technologies each demonstrate their own capabilities, known as NAM, covering chip organs, organoids, and computational models. Supporters believe that NAM may outperform animal models in simulating human biological characteristics and predicting the safety and efficacy of new drugs. Chip organs and organoids are typically constructed from human cells, and computational models can also be designed based on human data. For example, the "liver chip" developed by the Wise Institute for Bioinspired Engineering at Harvard University in the United States places human liver cells in microfluidic channels to simulate the blood flow environment and detect whether drugs may cause liver damage. Related studies have shown that this chip has an accuracy rate of 87% in identifying known hepatotoxic compounds, and has successfully identified multiple drugs that have passed animal experiments but are at risk of exposure in clinical stages. In 2024, this technology will be included in the FDA's innovative tool pilot program and is expected to serve as an alternative to some toxicity tests in the future. Organoids are another important route. Researchers have used induced pluripotent stem cells (iPSCs) to cultivate three-dimensional miniature "organs" for simulating disease progression and drug response. The Stanford University team proposed the concept of "clinical trials in petri dishes", which involves obtaining iPSCs from different patients, cultivating them into cells or organoids, and testing whether candidate drugs can improve the function of these "disease models". In a 2020 heart failure study, this method identified candidate drugs that are effective for patients with specific gene mutations. The significance lies in the hope of screening out ineffective or harmful schemes before entering animal or even human experiments. The third alternative method is computational modeling and artificial intelligence (AI), which simulates drug behavior through computers. In 2023, the FDA National Toxicology Research Center developed the birth form model "AnimalGAN", which uses existing animal data to generate "virtual rats" and predict drug toxicity performance. In the simulation experiment, the model successfully ranked the hepatotoxicity of structurally similar drugs. The goal of such tools is not simply to replicate animal experiments, but to reconstruct toxicology assessment logic using data. The application scale of alternative methods is expanding. Statistics show that between 2006 and 2022, the number of biomedical papers using only NAM has increased from approximately 25000 to 100000. Pharmaceutical companies have also begun to incorporate organoids and chip systems into their research and development processes in order to improve clinical success rates. The reason why animal models face difficulties in reality and alternative methods have received attention is closely related to the practical difficulties of animal models. In drug development, a large number of candidate drugs perform well in animal trials but fail in human clinical trials. Data shows that approximately 86% of candidate drugs ultimately fail clinical validation. Taking sepsis as an example, researchers have developed over 100 promising therapies in rodent models, but almost all have failed in human trials. One of the reasons is that there are significant differences in the immune systems between humans and rodents, and experimental mice usually have highly identical genes and a single living environment, making it difficult to simulate the complexity of real populations. In addition, animal testing itself is costly and time-consuming, and some studies also involve primates, causing ethical controversies. That is why in 2025, the UK will classify animal testing into three categories in its new policy: one category is rapidly replaceable, such as the upcoming cancellation of skin irritation testing this year, which will be replaced by computational, cellular, or chemical methods; The second type requires gradual reduction, such as partial pharmacokinetic studies that analyze the absorption, distribution, metabolism, and excretion processes of drugs in the body; The third category currently only includes one example, which is fish testing for endocrine disruptors in environmental monitoring, with the goal of developing alternative methods by 2035. The FDA has also released a roadmap to reduce, optimize, and replace animal testing in drug testing, with an initial focus on monoclonal antibody testing, as animal research is both expensive and difficult to predict human responses. NIH announced that it will no longer fund research projects specifically focused on animal models to encourage more adoption of NAM. However, alternative methods are not a 'master key'. Some researchers believe that in the foreseeable future, some animal studies will still be indispensable, such as complete organs with complex blood vessels and neural networks, interacting endocrine and reproductive systems, or biological processes such as tissue aging, which are currently difficult to replicate in organoids or chip organs. The alternative method of 'going live' should be approached with caution. Nature magazine points out that what truly determines the fate of NAM is not technical concepts, but 'validation'. Any model that wants to be used for drug or chemical approval must submit sufficient data to national or international validation agencies to demonstrate its accuracy, reproducibility, and generalizability. The verification paths of different methods vary greatly, often taking many years and being costly. Therefore, in order to accelerate the process, the UK plans to establish an alternative method validation center, and NIH has also invested tens of millions of dollars in the development of standardized organoid models. However, caution still exists on the front lines of regulation and scientific research. Some researchers are concerned that policy slogans may lead the public to mistakenly believe that alternative technologies have fully matured. In fact, the reasons for clinical failure of drugs are not only limited by animal model limitations, but also include trial design flaws, insufficient sample size, and other issues that may also occur in NAM studies. At present, most scientists tend to "advance in stages", which means rapidly phasing out animal testing in areas where mature alternative solutions already exist, gradually reducing it in other areas, and continuously investing in basic research. Animal testing may not completely disappear in the near future, but its' default status' is being shaken. From ethical pressure to technological progress, from policy commitments to regulatory games, this transformation is far from a simple "replacement" story, but rather a rebalancing of scientific reliability and human responsibility. The real challenge is just beginning when alternative methods are required to meet the same rigorous standards as animal models. (New Society)