In Science, Technology, Engineering and Maths (STEM) education, lab-based classes play an important role by allowing students to acquire experimental skills, as part of their training. This vital role is evidenced by a seminal paper authored by Feisel & Rosa (2005), which proposes a list of 13 fundamental objectives of instructional laboratories in undergraduate engineering degrees. This same paper also addresses the potential of combining different lab environments, like remote and virtual labs. The trend to expand the boundaries of traditional – hands-on – labs to the cyberspace, creating a new type of non-traditional – virtual and remote – labs has gained momentum with the COVID-19 pandemic. Nevertheless, this trend has roots that can be traced back to almost 30 years ago, as acknowledged by Froyd, Wankat & Smith (2012).

More recently, Artificial Intelligence (AI) has been impacting the entire educational landscape, with a particular emphasis on generative AI tools. This trend is clearly affecting enegineering education, although examples linked to the acquisition of experimental skills, specially through non-traditional labs, are still scarce. Noticeably, a recent example already showcases antiplagiarism and AI immunity features (Toh and Buchanan, 2025), which suggests AI being seem as a foe rather than as a friend.

The keynote will address this thematic, i.e., is AI a friend or a foe regarding traditional and non-traditional laboratories, or can this condition be controlled by a proper lab and/or instructional design?