Session Description
Artificial Intelligence (AI) is useful but has two separate downsides. Technically, AI has shortcomings. It can oversimplify. It can make mistakes large and small. It can generate grammatically correct text that does not offer useful results (‘workslop’). And, psychologically, many people fear losing work to it. This workshop offers positive tools to improve both situations. We show how to make more effective use of AI and how the human contribution is essential to a project.
We use the framework from the AI company Anthropic. They say people interact with AI in three ways: Automation, Augmentation and Agency. We work within the Augmentation framework—when AI works interactively as a creative partner.
We show that three tools, which help people convert technical data into complex and useful arguments, also enhance interaction with AI. The three are:
• Identify hierarchical and extended reasoning
• Build an argument outline
• Use relevant justifications to support conclusions
These argument skills help identify and overcome specific potential shortcomings with text prepared by AI. They help professionals identify essential elements in their project and how those elements support the step-by-step reasoning that extends from scoping through significance to mitigation. Mastering such specificity permits finding how AI has oversimplified steps, missed or incorrectly worded conclusions, or made errors in reasoning.
The workshop includes background information, specific examples, handouts and practice with realistic situations. Participants leave able to make powerful use of their expertise while collaborating with AI models.
Session Summary
The workshop shows how three tools of argument can help people guide and monitor the work of Artificial Intelligence. They help professionals use the powerful features of AI while maximizing their professional expertise to find and correct the errors and simplifications that AI can introduce.
Additional Information
Brown, G. (in press). Writing impact assessments with the goal of building strong arguments. In Fischer, T., Bice, S. Jha-Thakur, U. Montano, M. Noble, B. & Retief, F. (eds.) Impact Assessment Encyclopedia. Edward Elgar Publishers, Cheltenham, UK.
Batdorj, B. (in progress). Enhancing Mining Transparency in Sustainability Reporting through Artificial Intelligence: Theory, Evidence and Application. University of British Columbia, PhD dissertation.
Session Chair
Glenn Brown
Royal Roads University
Canada
Session Co-chair
Bulgan Batdorj
University of British Columbia
Canada