Experiment - figured out by machine

But machines can do even more. They can create plans of experiments based on rules of the field of the actual science. However those potential approaches require further supervision of researchers and management of the company whether to realize them or not, the way of creating new experiments with AI leads to a new level of the scientific creativity.

There is a difference between being a magnificent scientist and being creative. In fact those skills doesn’t depend on each other. To be successful on one or on other way needs various abilities or facilities. Artificial intelligence has limitless creativity within the given rules and human adds the experience what only researchers can achieve. They can work together to build a better method than they could build without collaboration.

An AI based system can invent hundreds of experiments in a couple of hours. Creative times definitely works much slower. With artificial intelligence you have to care about the validation and the realization of experiments only. Moreover when the AI processes the monotone, boring or overwhelming parts of work, the company could reallocate human resources to the field where scientist can be more productive and can achieve personal success without the fear of burn-out. Artificial intelligence has two more advantages, it can review every historical data of the actual field in a short time and it never forgets the successes and the failures of earlier experiments.

To build an experiment-generator-system we need only input data and a list of expected output. The AI can work from publicly accessible and in-company data as well. Defeat is impossible because the system learns from the opinion of the scientists. If they refuse to realize an experiment-suggestion, just let the system to know that and the reasons and it generates a couple of new experiments again. The AI teaches themself from every outcome.

Scientists works with generative networks in physics already. They know how to build better experiments. We know how to use this method in other field of science.

Final conclusion, if you are not satisfied with the quality and result of the experiments at your company, you might investigate how quality cats your researchers are drawing. But never forget, profitability and productivity over cuteness and fluffiness.