Beyond Generative A.I. to Reduce the Gap Between Architecture and Its Techniques

This work explores some methods for integrating Generative Artificial Intelligence (Gen.A.I.) into Technical Architecture processes. After the problem contextualization in AEC sector, including opportunities for enhancement, the paper analyses critically the state of art about A.I. development in the restricted field of study. In general we assume that, present and future outcomes both, in research and practice, depend on the pervasive hardware advancement for calculation power, together with the software usability simplification, based on intuitive middleware levels, up to natural language interfaces. Specifically this contribution aims to investigate how Gen.A.I. can become a component of project and construction methodology, capable of more efficient results, by reducing the gap between creative solutions and their feasibility. First, a text-to-image case of study is presented, aimed – rather than to assess the implementation pipeline – to illustrate the ease of the process, even more by using free software widely available online: it shows that the impressive visual result originates from an innovative methodology, produced not only by statistical knowledge, but also by artificial generative abilities. Follows a critical discussion about the potential of this “new” technology: it concerns not only aesthetic composition and form-finding, that seem to monopolise the attention of the sector, but clearly it involves the whole building process, including technique and technology of architecture. To this aim, authors select and analyse some A.I. applications properly addressing objectives of the BIM method, namely enhancing technical control on design solution performance. Conclusions outline the relevant role of Scientific Community, called to guide the transformation of these processes, by addressing structured Technical Architecture knowledge models, instead of letting it being guided by the automatic algorithms.