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 335

Mr Ali Khan

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 that errors in AI-generated submissions should be managed through disclosure and verification rather than prohibition, and that AI works as assistive technology for people with disabilities, neurodivergent people and non-native speakers. Says that the main danger is autonomous weapons and supports a binding international treaty.

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. Says AI-generated material contains errors such as hallucinated citations, which require verification by the author.

    AI-generated materials contain errors. Large language models produce hallucinated citations, misattribute research, and summarize studies incorrectly. Citations and legislation referenced by AI can all be erroneous, and they require verification by the author.

  2. Argues that when a matter before parliament is one of opinion, citation errors in a submission do not change the position advocated.

    When the matter before a parliament is fundamentally one of opinion, the factual error rate in a submission does not change the position being advocated.

  3. Argues that human error rates have been answered with regulation, training, insurance, oversight, disclosure and remedies rather than by prohibiting the underlying activity.

    We have never responded to human error rates by prohibiting the underlying activity. We responded with regulation, training, insurance, oversight, disclosure, and remedies.

  4. Argues that banning assistive AI discriminates in effect against disabled students who need it to achieve fairness.

    For many disabled students, banning AI does not restore fairness — it removes the only means by which fairness was ever achievable. Banning assistive AI discriminates, in effect, against the people who need it most.

  5. Argues that the correct target of fear is the weaponisation of AI, including autonomous machines deciding to kill and arms races among algorithms.

    The correct target of our genuine fear is the weaponization of AI — autonomous machines deciding to kill, arms races among algorithms, and the conceivable extermination of the human race by systems handed full military control.

Recommendations

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

  1. Impose restraints and bans on autonomous AI weapons on battlefields through the United Nations and responsible stakeholders.

    This is where genuine fear belongs, and where the United Nations and responsible stakeholders must impose restraints and bans on autonomous AI weapons on battlefields.

  2. Manage AI errors through verification duties, disclosure norms and remedies, not prohibition.

    Accept AI's error rate as we accept every other human-system error rate —through verification duties, disclosure norms, and remedies — not through prohibition and persecution.

  3. Preserve the value of opinion-based submissions, with parliament weighing the evidence and the citizen's support or opposition valid regardless of bibliographic imperfection.

    Preserve the dichotomous value of opinion-based submissions. Parliament weighs the evidence; the citizen's support or opposition is valid participation regardless of bibliographic imperfection.

  4. Normalise AI as assistive technology for disabled and neurodivergent people, non-native speakers and everyday productivity; protect access as an accessibility right.

    Normalize AI for people with disabilities, neurodivergent people, non-native speakers, and everyday productivity — as assistive technology, stabilizer, and equalizer — and protect access to it as an accessibility right.

  5. Focus institutional fear on autonomous weapons, systems making consequential decisions about humans without human judgment, and accountability for consequential AI harms, not on consumers typing documents or emails.

    Focus institutional fear where it belongs: on autonomous weapons, on systems making consequential decisions about humans without human judgment, and on accountability for consequential AI harms — not on consumers typing documents or emails.

  6. Adopt transparent norms: disclose AI use, verify citations, and hold writers and readers to validation duties.

    Adopt transparent norms: disclose AI use, verify citations, and hold both writers and readers to validation duties.

  7. Support the international push for a binding treaty on lethal autonomous weapons ahead of the November 2026 CCW Review Conference.

    Support the international push for a binding treaty on lethal autonomous weapons ahead of the November 2026 CCW Review Conference.

Read the original

Open in the interactive guide