Information technology is seeing a renaissance in the use of Artificial Intelligence (AI) in Education. While instructional delivery is using state-of the art multimedia, assessment of student work is becoming increasingly dependent upon normative analysis of small snippets of student work (e.g. multiple choice or short answer responses.) The mechanical grading systems currently employed encourage “disembodied” responses from learners, and do not support a classic Socratic approach where deep learning can take place. The Performance Assessment movement has demonstrated that portfolio construction of student initiated projects produces successful student learning outcomes. This workshop will introduce participants to a pedagogy and assessment methodology that grew out of the AI in Education movement of the 1980s. Students build portfolios using Google Sites for projects of their own design. The underlying data model, and interaction design facilitate iterative formative assessment, while simultaneously reducing the grading burden on the teacher.