Question 1

Would an Australian AI Act be bad for business?

Accepted Answer

No. It would provide clarity and support responsible AI use. Existing laws apply to AI, but how remains unclear. Model developers often offer black-box systems and reject liability, leaving businesses uncertain of their obligations. This challenge intensifies as government and market pressures push adoption.

Other sectors have established laws and norms. Car makers meet safety standards, businesses maintain fleets and provide training, and drivers obey road rules. When something goes wrong, responsibility is typically clear. In AI, there is little guidance and limited specialist support. Litigation has already started overseas, including a case where a chatbot caused serious harm. Australian businesses deploying AI may soon face similar litigation.

How an AI Act would help:

Regulator clarity: Require existing regulators to clarify how their rules apply to AI and issue timely guidance.
Defined responsibilities: Outline obligations for developers, deployers and end users, preventing liability offloading or at minimum providing certainty.
Guardrails and standards: Introduce practical technical and operational requirements for safe AI use, helping businesses manage risks.
Liability frameworks: Clarify accountability for harms, reducing legal uncertainty for Australian businesses.
Transparency obligations: Require developers to publish risk evaluations and detailed reports on model capabilities and limitations, helping businesses understand what they're buying.
Harmonisation: Align Australian AI rules with the EU to streamline global operations and enhance trade.

Question 2

Are you saying that all AI systems are dangerous?

Accepted Answer

No. Today's LLMs are unlikely to pose catastrophic or existential risks. While there are valid ethical concerns, today's AI systems offer important economic opportunities while being unlikely to pose extreme safety risks. However, evaluations by leading labs and third parties show an increasing likelihood that LLMs will soon pose a range of risks, like being able to assist wrongdoers build chemical, biological, radiological and nuclear weapons, conduct sophisticated cyber attacks at scale, or even escape human control. Given AI scaling laws, the need to prepare for highly capable AI systems that could be dangerous has become pressing.

Question 3

What do you mean by catastrophic or existential risk?

Accepted Answer

Without providing a technical definition, we mean that AI could pose a risk of a similar scale to pandemics or nuclear war. This could include social collapse or large numbers of deaths. The ability to assist wrongdoers build chemical, biological, radiological and nuclear weapons is one way this could occur, and is currently a focus area of researchers and other governments. A taxonomy of AI risks, including extreme risks, is available in the MIT AI Risk Repository (including 7.2).

Question 4

Who is behind Australians for AI Safety?

Accepted Answer

Australians for AI Safety is volunteer-run, has received no funding from any sources, and doesn’t ask for or accept funding from its supporters. Our supporters share common values, most importantly, we want AI to be safe, in Australia and globally. Our thanks go to Good Ancestors who play a coordinating role, provide technical support, policy development, support and host our website.

Question 5

Do the signatories of the open letter endorse the footnotes and website content?

Accepted Answer

When someone signs any of our open letters they are only signing onto the core text of the letter. The rest of the website including the footnotes on the letter do not necessarily represent the views of signatories of the open letters.

Question 6

How is my data handled when I sign the open letter?

Accepted Answer

When you create a an account or sign the open letter, we collect only the information necessary to verify your signature. Your email address is used for verification and optional updates. We do not share your personal information with third parties without your consent. You can request removal of your information at any time. The handling of your personal information covered by the Good Ancestors privacy policy.

Frequently Asked Questions

AI Policy

Would an Australian AI Act be bad for business?

Are you saying that all AI systems are dangerous?

What do you mean by catastrophic or existential risk?

About Us

Who is behind Australians for AI Safety?

Open Letters

Do the signatories of the open letter endorse the footnotes and website content?

Privacy

How is my data handled when I sign the open letter?

Still have questions?