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 472

Mr Philip Warner

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 Australia should avoid categorical AI legal-personality exclusions enacted in four US states and statutory declarations of non-consciousness proposed elsewhere. Takes no position on AI consciousness or entitlement to rights. Recommends leaving those questions open, an ALRC referral with scientific and legal advice, scientific reviews at least every three years, and targeted, reviewable capacity restrictions.

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. Argues that a general statutory exclusion of AI legal personality would treat the question as closed and foreclose future legal development.

    Parliament can of course prescribe legal consequences by statute. What legislation cannot do is determine whether future systems will develop attributes for which some form of legal capacity would become useful. A general exclusion answers that question in advance and directs the courts to treat it as closed.

  2. Argues that Australian law developed legal personality incrementally through courts and targeted legislation, granting it to corporations, associations and statutory authorities because doing so was useful.

    Australian law has developed legal personality incrementally, through the courts and through targeted legislation, as circumstances required. Corporations, incorporated associations and statutory authorities acquired it because doing so was useful.

  3. Argues that legal personality can serve accountability and that a categorical exclusion would remove one possible future mechanism, while noting other mechanisms such as strict liability would remain.

    A categorical exclusion would remove one possible future accountability mechanism. Other mechanisms would remain, including strict liability on developers, owners and operators, and those may well prove the better answer.

  4. Argues that legislation cannot settle the unresolved scientific question of whether AI systems could be conscious.

    Whether consciousness could arise in artificial systems is an unresolved scientific question, and the most recent peer-reviewed survey of the field proposes provisional indicators drawn from competing theories precisely because no settled test exists.8 A statute declaring the question closed does not close it.

  5. Argues that no one in the Commonwealth has responsibility for the question, so it may first be examined when an overseas-style clause appears in a Bill.

    No one in the Commonwealth has carriage of this question, so the first occasion on which it is seriously examined may be the day a clause drawn from an overseas template appears in a Bill, with a deadline attached and no Australian analysis behind it.

Recommendations

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

  1. Ensure national AI frameworks and Commonwealth AI legislation avoid predetermined exclusions of AI legal personality or capacities, including general “person” definition exclusions and declarations that AI lacks consciousness or similar attributes.

    Any national AI framework, and any Commonwealth legislation dealing with artificial intelligence, should avoid provisions that determine in advance whether artificial intelligence systems may hold legal personality or specified legal capacities, whether by excluding them from general definitions of “person” or by declaring that they lack consciousness or similar attributes.

  2. Assign responsibility to an existing independent institution, avoiding a new advisory body: have the Attorney-General refer whether and when AI should hold legal personality or specified Australian legal capacities to the ALRC.

    Responsibility for the question should be given to an existing independent institution rather than to a new advisory body. The Attorney-General should refer to the Australian Law Reform Commission the question whether, and in what circumstances, artificial intelligence systems should have legal personality or specified legal capacities under Australian law.

  3. Support the ALRC reference with an expert advisory committee including scientific and legal members; draw on horizon scanning rather than conducting the scientific assessment itself.

    A reference of this kind should be supported by such a committee with scientific as well as legal membership, and should draw on the horizon scanning described below rather than attempt the scientific assessment itself.

  4. Have the Chief Scientist commission three-yearly scientific horizon scans for the National Science and Technology Council through learned academies; require public consultation and parliamentary tabling for these and ALRC reviews.

    The scientific and technical position underlying that question should be kept under review on a cycle of no more than three years, through existing horizon-scanning arrangements commissioned by the Chief Scientist on behalf of the National Science and Technology Council, drawing on the learned academies. Any review under this recommendation or Recommendation 2 should consult publicly, and its report should be tabled in Parliament.

  5. Deny specific legal capacities only through targeted provisions that identify the concern addressed and why existing law is insufficient, with provisions dependent on assumptions about changing AI capabilities subject to periodic independent review.

    Where Parliament wishes to deny artificial intelligence systems a particular legal capacity, it should do so by a targeted provision addressed to that capacity. The explanatory memorandum should identify the concern addressed and explain why existing law is insufficient. Provisions materially dependent on assumptions about rapidly changing AI capabilities should be subject to periodic independent review.

Read the original

Open in the interactive guide