Despite research and developement in the field of Artificial Intelligence (AI) has been conducted for decades, the past months showed arapid change in the general perception of the topic in the general public. This had to do - amongst other aspects - with the appearance andgeneral roleout of tools and instrumentes easy-to-use text processors and large language models, such as Chat GPT or Bard. A lot ofdiscussion about the potential of supporting or even replacing AEC (Architecture - Engineering - Construction) professionals can bemonitored in recent days. For the AEC domain it seems important that support of planners and engineers is reliable. As such, thiscontribution targets the question, if - based on empiric exeriments - AI-instruments already provide in-depth reliable encyclopaedic andknowledge-based answers on requests pertaining to the building physics domain. Given that the "AI revolution" might end up in new,everyday-usage common tools, as the did the WWW or even pocket calculators with equation solving capabilities before, reliability andvalidity of answers is of crucial importance. The paper containts approach, method, empiric experiments, results, and an in-depth discussionabout the current state of AI as a consultant and tutor for AEC-related questions.