AI Inquiry Atlas

An unofficial, independent guide to the public records of the Joint Select Committee on Artificial Intelligence. Not connected with the Parliament of Australia or the committee. Its makers also made a submission (no. 481; see About). Summaries and scores are written by AI and can be wrong: check the original before relying on them.

Submission 388

Mr Nathan Lou

Submitter group: Individuals.

Position labels are not shown for individual submitters.

An AI model identified the submitter group and wrote the summary, key points and recommendation summaries. Counts do not measure importance or merit.

Open in the interactive guide

Summary

Read the original submission for the full text.

Argues from a public key infrastructure background that AI companies lack independently verified safety assurance, and that current legislation is insufficient. Recommends mandatory reporting and audits, a stronger AI Safety Institute and whistleblower protections, and discusses sovereign AI and the electronics industry.

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.

  1. Asks what is stopping AI companies from crippling Australia.

    As a concerned citizen, I only ask one simple question: “What’s stopping AI companies from crippling Australia?”

  2. Says that, as far as the author knows, no AI company has had its adherence to its safety policies independently audited.

    In comparison with AI companies, some choose to volunteer safety policy documents, but none as far as I know have themselves audited by an independent party to verify their own adherence to such policies.

  3. Says that in the author's professional opinion it is difficult to trust AI companies to prioritise safety over profit under current legislation, and that the best trust is one that is not needed.

    Thus, it is in my professional opinion, that it would be quite difficult to trust AI companies to keep AI safety a priority over profit, with the legislation in place as it is now. From my experience, the best type of trust, is one that is not needed.

  4. Describes a Kevin Rudd speech urging sovereign AI across the AI stack, and says in the author's opinion this echoes mutually assured destruction logic.

    In a recent speech from Kevin Rudd, he pushed for the urgent development of sovereign AI, across the entire AI stack. In my opinion, this sentiment echoes MAD logic, wherein the only defence against a weapon of mass destruction, is one of your own.

  5. Argues that delivering truly sovereign AI would require reviving Australia's sovereign electronics industry.

    in order to deliver a truly sovereign AI, we will need to revive our sovereign electronics industry, and this may be our one generational chance to do so, or to lose it forever.

  6. Argues that due process must not be skipped in AI development, warning otherwise of a 'Chernobyl'.

    All in all, as Australia finds itself in this unique opportunity to not only ride the wave of success, but potentially direct where it leads, due process can not, and must not be skipped – or else we’ll have a Chernobyl on our hands.

Recommendations

Recommendations are shown beside the submission’s own words. Listed in the same order as on its page in the interactive guide.

  1. Mandate standardised reporting of AI safety concerns and events, for example categorised into deceptive behaviour, escapement and miscellaneous.

    Mandatory reporting of AI safety concerns/events, in a standardised fashion E.g categorised into deceptive behaviour, escapement, misc., etc.

  2. Mandate audits of AI companies' safety policies by third parties.

    Mandatory audits of safety policies by 3 rd parties

  3. Strengthen whistleblower protections.

    Similarly, I support: - Mandatory reporting of AI safety concerns/events, in a standardised fashion E.g categorised into deceptive behaviour, escapement, misc., etc. - Mandatory audits of safety policies by 3 rd parties - Stronger whistleblower protections

  4. Empower the AI Safety Institute, similar to the OAIC.

    Empowering the AI safety Institute – similar to the OAIC

  5. Engage internationally for consolidated controls.

    International engagement for consolidated controls

  6. Impose a duty-of-care responsibility on AI companies.

    A Duty-of-care responsibility to AI companies

  7. Induce industry to provide sovereign AI components.

    I think there are great benefits to be had by inducing industry to provide sovereign AI components.

  8. Incentivise companies at the forefront of AI to reverse sovereign capability loss through a genuine paradigm shift.

    Institutional knowledge loss is eating away at Australia’s sovereign capabilities, and without clear direction from the government, or the market, is destined to wither away. By incentivising companies to be at the forefront of the AI boom, we can reverse this current trend. But it needs to be a genuine paradigm shift.

Read the original

Open in the interactive guide