I believe that experiments should be duplicated under the same conditions to check if results are correct and genuine.  Besides that experienced reviewers will be able to gauge if the results are feasible and correct.  However, anti plagiarism software can only detect similarity index and plagiarism, not the data itself.  Perhaps if the data is collected by a group of people, there may be a greater chance that it's genuine.  Thanks.  Have a good day.

Should we trust the research results of a researcher who has no ethics in real life?

Think we should give the person "the benefit of the doubt" i.e. s/he is telling the truth until proven otherwise and can't generalize his / her research results based on his / her personal characters.  We still need to be prudent in reviewing / evaluating his / her research work end to end to make a sound judgement.

Plagiarism constitutes a very tiny part of the many types of frauds and scams that can be committed by a dishonest scientist, don’t you agree? What I meant in the original question is a little bit different. Suppose you have a colleague that is known to be a professional liar in daily life and you know that he/she would do and say anything to achieve even minor benefits. When such a colleague publishes his/her work and claims that his/her results are correct and were achieved after conducting his/her experiment under proper conditions, should we trust that? This is the question.