Skip to main content
TPDI logo
Back to Layer 3

Capability 3.1.1.1

Computer Vision

Definition

Models that interpret visual inputs (images, video, sensor data) for detection, classification or understanding.

Examples

  • Harrison.rad.1 (Harrison.ai)
  • YOLOv8 (Ultralytics)
  • Vision Transformer (ViT)
  • CLIP (OpenAI) Nearmap AI
  • DroneShield (DroneOpdID)
  • iBenthos
  • EMUSE.

Indicators

  • Number of domestically developed models
  • Number of papers published on model development by researchers working for Australian entities

National AI Plan

PLAN COMMITMENT

Limited commitment

General: The government is considering the role of Australian-developed models. References $460 million in existing funding already available or committed to AI and related initiatives; notes existing and general $1 billion commitment for critical technologies in the national interest under the National Reconstruction Fund; notes applicability of Research and Development Tax Incentive Program to AI (Action 2) as well as intention to attract investment (Action 3) and for government to invest in sovereign AI for public service (Action 3), but no specific programs for model development in general, nor this specific category. This capability: No specific focus, but the plan recognises the value of AI in agriculture, resources and environmental monitoring, where computer vision capability is relevant. Limited commitment by government in the plan represents a missed opportunity. More should be done to leverage this very high agency. This is one of 8 capabilities over which Australia wields very high advantage; national priority should be given to maintaining this agency and leveraging it internationally to offset gaps in other capabilities of low agency.

Stocktake

Is the Stocktake missing evidence?

Share additional information for potential future Assessments