Generative AI and large language models have already changed many areas of society, and education is not an exception. In this paper, we present a study where 302 higher education teachers were surveyed about their experiences, attitudes, and ethical perceptions towards AI in education. The first instance of the survey (N=177) was conducted in late spring 2024, and the second one (N=125) less than a year later, with a focus on technical universities. According to the results, the teachers in general do not really use AI tools in everyday education, although there are a few exceptions who use them quite a lot. The main reason for not using AI may be a lack of proper training, as it seems that the universities either do not offer such training or teachers do not participate in it. Regardless, teachers find teaching AI to students very important, but also seem to agree that the use of AI tools has increased cheating. Moreover, the teachers find the ethical aspects and data security very important when AI is used. In the paper, we provide a detailed look at the results, with focus on ethical aspects, and offer recommendations based on the analysis.