The AI can be tested for poison, will the animals leave the lab?
According to statistics, in Europe alone, the number of animals used for experiments is more than 10 million per year. Globally, it may break billions.
Anatomy, toxicology experiments, drug experiments... It can be said that animals have made great sacrifices for the advancement of human medical research. Especially the white mice, which often show red eye characteristics due to illness, the number of deaths per year is even more difficult to count.
In addition, many experiments do not usually anesthetize the animals, or only a small dose, in order to avoid the effects of the drug on the actual test results. The pain caused by the animals is unimaginable.
Therefore, many medical freshmen who are enrolled are struggling in the face of their first anatomy class...
However, due to the particularity of biological experiments, it is impossible for humans to conduct research on their own bodies, although this is the most reasonable way. Then, these pains have to be endured by animals that are more similar to a certain physiological feature of humans. But even so, the results and significance of experiments with animals are limited.
In those years, the animals suffered injuries in the laboratory.
The use of animals as specimens for medical experiments for humans is itself the biggest paradox. Although they are all creatures, the differences in individual characteristics are destined to have only a reference value for the data obtained in animals. This is why many drugs have to undergo a human testing phase before they are officially placed on the market.
Of course, it can enter the human testing phase, indicating that it has great security.
But the pain point of animal experiments is far more than that.
The first is the risk of life in experimental animals. Humans choose to use animals instead of themselves to do experiments, the purpose is to avoid the possible risk of death. But this does not mean that animal life can be arbitrarily ruined. In the actual operation process, due to a series of human factors such as experimental design and operational errors, the number of accidental deaths of experimental animals is large. Under the demand of scientific research, such accidental death is considered reasonable. But from the point of view of experimentation rather than life, this is a huge waste of resources.
Secondly, the cycle of animal experiments is longer. Depending on the purpose of the experiment, injection and administration of the animal will often last for a longer period of time, during which time more effort is required to record and observe. The purpose is to observe the natural tolerance response of the animal. This is also why the development of drugs often takes a period of ten years or more. For many patients, the speed of drug development is often not as fast as the disease.
Finally, the repeatability of animal experiments. For example, the Dreiser test, which was originally used in the toxicity test of cosmetics. Because the toxicity of the test article is unknown, it needs to be administered to the animal over and over again. Some drugs can even be tested six or seventy times, which is simply a thousand knives. This process will undoubtedly cause more pain to the animals.
That is to say, the animal experiment is not completely suitable for the limitations of the human body. It still has the objective reality that has to be faced above. The added pain to the animals is self-evident.
Therefore, the 3R principles of animal experiments (replancement, reduction, refinement) advocate the use of unconscious materials instead of animals, improve the accuracy of experiments in various ways to reduce animal use and optimize animal signs to achieve relatively accurate experimental data. The best way is of course to replace animal experiments, but under today's conditions, this is undoubtedly an impossible thing.
In fact, scientists are gradually replacing animal experiments, such as using cell experiments.
And as technology advances, scientists begin to use another secret weapon: machine learning. Animals will benefit from this at least in terms of testing toxicity.
The test is no longer an animal, but an AI.
Recently, a study from the Journal of Toxicology showed that the history of using animal testing toxicity through a new machine learning system is likely to be rewritten. Animal experiments, at least animal testing toxicity tests, may be replaced by computers.
The research team began to accumulate data from 2014 and eventually collected 10,000 chemicals. The AI system was then used to test 860,000 chemical substances in the database, which mapped out previously unknown relationships between molecular structure and specific toxicity types, such as their possible effects on the eyes, skin, or DNA.
Specifically, the researchers grouped substances of different chemical and toxicological properties, and then speculated on the toxicity of other substances based on the toxicity of a substance in the group. This is in contrast to the known toxicity of these other substances in the database, resulting in computer predictions.
The final results showed that the accuracy of the toxicity obtained by computer testing was 87%, while the correct rate of testing with animals was 81%. That is to say, in terms of toxicity testing, AI has already won the animal. Based on this data alone, AI can already achieve animal replacement.
Of course, it is a bit too early to say that the alternative is too early, especially to replace animal experiments.
For example, data problems. As we mentioned above, it is only for such a simple toxicity test that the research team spent four years on data accumulation and experimental analysis. It is not difficult to imagine that if you are doing more kinds of toxicity tests and more difficult disease drug research, the data that you need to collect will inevitably be more or less. Then, how to build such a huge database, while controlling the construction time, will probably be a project of ten years.
On the other hand, animals themselves have some advantages that computers are hard to match. The computer works on the basis of known data, and the animal may have some unknown, possibly unexpected, gains in the course of the experiment. For example, many physical or biological science discoveries are derived from an accident. So, is the computer bound by a fixed program, and is it possible to have some magical performance that can happen to animals?
Maybe, this also requires an additional computer simulation system.
The protagonist and supporting role of biological experiment
Admittedly, this is just a small step in the field of biological experiments. This small step also shows us the positive significance of its biological experiments.
1. Strengthening and weakening the 3R principle. The 3R principle is based on the idea of respecting animal life. Among the three, if the "replacement" is achieved, there will be no conditions for the latter two. Computers are an important factor in strengthening "alternatives." If computer simulation analysis can be used instead of animal tolerance experiments, a significant number of animals will be replaced, the animal's natural survival rights will be guaranteed, and the latter two "R" concerns may be weakened.
2. The cost of biomedical research is greatly reduced. On the one hand, from the complicated affairs of breeding and managing animals, a computer system can adapt to all experiments, thus eliminating the cost of biological breeding. In comparison, the cost of computer maintenance is much lower. On the other hand, the experimental conclusions drawn by the computer are much faster than the natural response of the observed animals, and the cost of time will be greatly reduced, while also improving the efficiency of the experiment. For example, a lot of data has been experimentally obtained, but in new experiments, it will be repeated for various reasons, resulting in waste of resources and time.
3. The accuracy of the experiment is greatly improved. Based on a large and rich database, computers can make a large number of combined allocations of data resources to arrive at an optimal conclusion. This is hard to match with animal experiments. Based on the reaction of several mice and the combined analysis of 10,000 times or more, the accuracy is not the same.
But in reality, animals are, after all, the closest to humans in some respects, such as mice that have been shown to have very similar genetic sequences to humans. According to the current level of computer development, it is still difficult to achieve complete technical simulation. From this point of view, even if the computer is used as the protagonist of the bio-experimental technology, the animal entity may still be absent from the experiment as a supporting role. And what we have to do is to let the animals bear less sacrifice and pain. This is both an inevitable choice for technological progress and an ardent expectation of moral ethics.
Follow Me
——END——
Anatomy, toxicology experiments, drug experiments... It can be said that animals have made great sacrifices for the advancement of human medical research. Especially the white mice, which often show red eye characteristics due to illness, the number of deaths per year is even more difficult to count.
In addition, many experiments do not usually anesthetize the animals, or only a small dose, in order to avoid the effects of the drug on the actual test results. The pain caused by the animals is unimaginable.
Therefore, many medical freshmen who are enrolled are struggling in the face of their first anatomy class...
However, due to the particularity of biological experiments, it is impossible for humans to conduct research on their own bodies, although this is the most reasonable way. Then, these pains have to be endured by animals that are more similar to a certain physiological feature of humans. But even so, the results and significance of experiments with animals are limited.
In those years, the animals suffered injuries in the laboratory.
The use of animals as specimens for medical experiments for humans is itself the biggest paradox. Although they are all creatures, the differences in individual characteristics are destined to have only a reference value for the data obtained in animals. This is why many drugs have to undergo a human testing phase before they are officially placed on the market.
Of course, it can enter the human testing phase, indicating that it has great security.
But the pain point of animal experiments is far more than that.
The first is the risk of life in experimental animals. Humans choose to use animals instead of themselves to do experiments, the purpose is to avoid the possible risk of death. But this does not mean that animal life can be arbitrarily ruined. In the actual operation process, due to a series of human factors such as experimental design and operational errors, the number of accidental deaths of experimental animals is large. Under the demand of scientific research, such accidental death is considered reasonable. But from the point of view of experimentation rather than life, this is a huge waste of resources.
Secondly, the cycle of animal experiments is longer. Depending on the purpose of the experiment, injection and administration of the animal will often last for a longer period of time, during which time more effort is required to record and observe. The purpose is to observe the natural tolerance response of the animal. This is also why the development of drugs often takes a period of ten years or more. For many patients, the speed of drug development is often not as fast as the disease.
Finally, the repeatability of animal experiments. For example, the Dreiser test, which was originally used in the toxicity test of cosmetics. Because the toxicity of the test article is unknown, it needs to be administered to the animal over and over again. Some drugs can even be tested six or seventy times, which is simply a thousand knives. This process will undoubtedly cause more pain to the animals.
That is to say, the animal experiment is not completely suitable for the limitations of the human body. It still has the objective reality that has to be faced above. The added pain to the animals is self-evident.
Therefore, the 3R principles of animal experiments (replancement, reduction, refinement) advocate the use of unconscious materials instead of animals, improve the accuracy of experiments in various ways to reduce animal use and optimize animal signs to achieve relatively accurate experimental data. The best way is of course to replace animal experiments, but under today's conditions, this is undoubtedly an impossible thing.
In fact, scientists are gradually replacing animal experiments, such as using cell experiments.
And as technology advances, scientists begin to use another secret weapon: machine learning. Animals will benefit from this at least in terms of testing toxicity.
The test is no longer an animal, but an AI.
Recently, a study from the Journal of Toxicology showed that the history of using animal testing toxicity through a new machine learning system is likely to be rewritten. Animal experiments, at least animal testing toxicity tests, may be replaced by computers.
The research team began to accumulate data from 2014 and eventually collected 10,000 chemicals. The AI system was then used to test 860,000 chemical substances in the database, which mapped out previously unknown relationships between molecular structure and specific toxicity types, such as their possible effects on the eyes, skin, or DNA.
Specifically, the researchers grouped substances of different chemical and toxicological properties, and then speculated on the toxicity of other substances based on the toxicity of a substance in the group. This is in contrast to the known toxicity of these other substances in the database, resulting in computer predictions.
The final results showed that the accuracy of the toxicity obtained by computer testing was 87%, while the correct rate of testing with animals was 81%. That is to say, in terms of toxicity testing, AI has already won the animal. Based on this data alone, AI can already achieve animal replacement.
Of course, it is a bit too early to say that the alternative is too early, especially to replace animal experiments.
For example, data problems. As we mentioned above, it is only for such a simple toxicity test that the research team spent four years on data accumulation and experimental analysis. It is not difficult to imagine that if you are doing more kinds of toxicity tests and more difficult disease drug research, the data that you need to collect will inevitably be more or less. Then, how to build such a huge database, while controlling the construction time, will probably be a project of ten years.
On the other hand, animals themselves have some advantages that computers are hard to match. The computer works on the basis of known data, and the animal may have some unknown, possibly unexpected, gains in the course of the experiment. For example, many physical or biological science discoveries are derived from an accident. So, is the computer bound by a fixed program, and is it possible to have some magical performance that can happen to animals?
Maybe, this also requires an additional computer simulation system.
The protagonist and supporting role of biological experiment
Admittedly, this is just a small step in the field of biological experiments. This small step also shows us the positive significance of its biological experiments.
1. Strengthening and weakening the 3R principle. The 3R principle is based on the idea of respecting animal life. Among the three, if the "replacement" is achieved, there will be no conditions for the latter two. Computers are an important factor in strengthening "alternatives." If computer simulation analysis can be used instead of animal tolerance experiments, a significant number of animals will be replaced, the animal's natural survival rights will be guaranteed, and the latter two "R" concerns may be weakened.
2. The cost of biomedical research is greatly reduced. On the one hand, from the complicated affairs of breeding and managing animals, a computer system can adapt to all experiments, thus eliminating the cost of biological breeding. In comparison, the cost of computer maintenance is much lower. On the other hand, the experimental conclusions drawn by the computer are much faster than the natural response of the observed animals, and the cost of time will be greatly reduced, while also improving the efficiency of the experiment. For example, a lot of data has been experimentally obtained, but in new experiments, it will be repeated for various reasons, resulting in waste of resources and time.
3. The accuracy of the experiment is greatly improved. Based on a large and rich database, computers can make a large number of combined allocations of data resources to arrive at an optimal conclusion. This is hard to match with animal experiments. Based on the reaction of several mice and the combined analysis of 10,000 times or more, the accuracy is not the same.
But in reality, animals are, after all, the closest to humans in some respects, such as mice that have been shown to have very similar genetic sequences to humans. According to the current level of computer development, it is still difficult to achieve complete technical simulation. From this point of view, even if the computer is used as the protagonist of the bio-experimental technology, the animal entity may still be absent from the experiment as a supporting role. And what we have to do is to let the animals bear less sacrifice and pain. This is both an inevitable choice for technological progress and an ardent expectation of moral ethics.
Follow Me
Link:Tenco
——END——
评论
发表评论