Submission 134
Minderoo Foundation
Submitter group: Civil society.
An AI model identified the submitter group and wrote the summary, key points and recommendation summaries. Counts do not measure importance or merit.
Summary
Read the original submission for the full text.
Minderoo’s SafeAI proposes safeguards and sovereign capability through a Digital Duty of Care, privacy reform, an AI Safety Institute, copyright clarity, compute strategy, testing and Indo-Pacific leadership. Its attached roundtable discusses civil remedies, enforcement and international coordination. Its copyright project covers training, excluding inference and outputs; it presents eight options without selecting one, with the full report pending.
Key points
Quotes are the submitter’s own words. Text taken from documents, including scanned pages read by text-recognition software, can differ from the PDF. Check the original. Listed in the same order as on the submission’s page in the interactive guide.
States that SafeAI's work points to seven practical actions to build public trust, strengthen sovereign capability and capture AI's benefits.
SafeAI’s work points to seven practical actions that can build public trust, strengthen Australia’s sovereign capability and help the country capture the benefits of AI.
Describes survey findings that almost two-thirds believe AI is moving too fast and 46 per cent say risks outweigh benefits, while most prefer balanced rules that still allow innovation.
Almost two-thirds of those surveyed believe AI is moving too fast and 46 per cent say the risks outweigh the benefits. At the same time, most prefer a balanced approach with clear rules that protect people while still allowing innovation
Says AI developers need access to high-quality data, much of it copyrighted, and that creators and rights holders need clear and workable rules on use of their work.
AI developers need access to high-quality data, much of which is protected by copyright. Creators and rights holders need clear and workable rules governing how their work can be used and how they can benefit.
States that AI can be used to generate manipulated sexual imagery, support impersonation and fraud, and facilitate stalking, at greater speed and scale.
AI can be used to generate manipulated sexual imagery, support impersonation and fraud, facilitate stalking and harassment, and increase the speed and scale at which harmful material can be created and distributed, whilst also embedding and reproducing gender bias at scale.
Says that the roundtable it convened identified AI testing and evaluation as an area where Australia could build sovereign capability, especially for higher-risk uses.
The May roundtable identified AI testing and evaluation as an area where Australia could build sovereign capability, helping ensure AI models align with Australian laws and expectations before deployment, particularly in higher-risk settings.
Describes roundtable discussion of a civil right of action and accountability for large digital platforms, including AI-powered chatbots.
Participants discussed building on the reforms already underway, the digital duty of care andthe Scams Prevention Framework’s dispute resolution scheme, to ensure Australians would have a civil right of action and platforms are held accountable when duties are breached by large digital platforms, including AI-powered services like chatbots.
A program corrected obvious text-recognition errors in this quote for display.
Recommendations
Recommendations are shown beside the submission’s own words. Listed in the same order as on its page in the interactive guide.
Implement a Digital Duty of Care requiring technology companies to build safety into products, tools and services, with platform accountability and harm prevention.
Minderoo welcomes the Government's release of the Digital Duty of Care Exposure Draft. By mandating the tech sector to hard-wire safety into product design and delivery, governments can create a baseline and hold companies to account. The proposed duty would require technology companies to build safety into the design and delivery of their products, tools and services, creating a clearer baseline for accountability, harm prevention and ensuring their platforms do not facilitate abuse.
Continue Privacy Act reform.
2. Continue Privacy Act reform
Establish and properly resource the AI Safety Institute.
Establish and properly resource the AI Safety Institute
Provide regulatory clarity on the use of copyright material in AI training.
Provide regulatory clarity on the use of copyright material in AI training
Develop a data centre and compute strategy anchored in energy.
Develop a data centre and compute strategy anchored in energy
Build national AI testing and evaluation capability.
Build national AI testing and evaluation capability
Strengthen Australia's leadership on AI governance in the Indo-Pacific.
Strengthen Australia’s leadership on AI governance in the Indo-Pacific
Read the original
In some quotes a program corrected obvious text-recognition errors for display.