Author: Trusted Autonomous SystemsPublished: 22/12/2022Category: Standards, frameworks, and guidelinesLast updated: 04/04/2023

Autonomous Systems Demonstration Canvas

Cite as: Hanly, G., Wensing, E., McKerral, A., Donald, T., Ng, Simon, Scholz, J. and Devitt, S.K. (2022). Autonomous Systems Demonstration Canvas. Trusted Autonomous Systems. Toowong, Australia

The Autonomous Systems Demonstration Canvas ‘Demonstration Canvas’ is designed to help optimise human understanding and buy-in for technology developers and investors during demonstration of autonomous systems and AI. The Canvas provides a scaffolding framework to help technology developers efficiently prepare, practise, and perform impactful technology demonstrations to diverse stakeholders to achieve diverse goals including attracting investment, showing technical progress, and getting publicity.

Demonstrating the viability of artificial intelligence (AI) requires thoughtful construction and communication of both social and technical aspects. The Autonomous Systems Demonstration Canvas to help optimise human understanding and buy-in for technology developers and investors. The Canvas will help tackle the opacity problem facing AI, for instance the way AI can obfuscate the rules, reasoning, and justifications underlying human-machine decision making as it occurs.

Autonomous Systems Demonstration Canvas Handbook (.docx, .pdf)

Resource – S1 – Audience Persona Template (.pdf)

Resource – S1 – Roles by Canvas Section Map (.docx)

Resource – S2 – Demonstration Scenario Storyboard (.docx)

Resource – S2 – Explainable AI Handout (.docx)

Resource – S3 – Integrated Run-Sheet Template (.docx)

Resource – S4 – Demonstration Delivery Tips (.docx)

TAS_Canvas_WORKBOOK (.xlxs)

The Autonomous Systems Demonstration Canvas (‘Canvas’) optimises human understanding and buy-in for RAS-AI technology users and investors during demonstrations and provides a process to help plan, run and evaluate demonstrations across teams.

The Canvas and resources provide a framework, information, and guidance for building trust between demonstrators, their audience, and AI machines, helping to tackle the opacity problem facing AI, such as:

• the unpredictability of exactly how an autonomous system will decide to reach an objective.
• the challenge of explaining and justifying the underlying human-machine decision making as it occurs.
  common biases and flaws of logic in human perception and decision-making.
• the complexity of adapting legal, regulatory, and assurance requirements to RAS-AI technologies and exemplifying how they were met autonomously.

By providing resources that develop the skills and capabilities of demonstrators and the quality of their events, TAS can help facilitate transparent communication between industry and Defence, accelerating the iteration and acquisition paths of RAS-AI systems that is good for industry and good for Defence

See also, Devitt, S.K. (2022, 8 June). How to Plan a Successful Autonomous Systems Demonstration

https://tasdcrc.com.au/how-to-plan-a-successful-autonomous-systems-demonstration/.

The first iteration of the ‘Autonomous Systems Demonstration Canvas’ V.1.1 (ASDC) launched at the Queensland Defence Science Alliance (QDSA) Human Teaming and Response Robotics Standardised Testing, Evaluation and Certification Interactive Showcase at CSIRO’s Queensland Centre for Advanced Technologies’ MILTECS facilities on 21 April 2022