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 393

Name withheld

Submitter group: Individuals.

Position labels are not shown for individual submitters.

An AI model identified the submitter group and chose the extracts. Counts do not measure importance or merit.

Open in the interactive guide

Extracts from the submission, in the submitter’s own words.

Extracts

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. My concern is not that we should stop AI development. It is that as AI capabilities increase, the safeguards and independent oversight surrounding them should increase too.

  2. Current AI systems are not capable of independently taking control from humans. However, the technology is developing incredibly quickly.

  3. The greater the potential consequence if something goes wrong, the stronger the evidence and independent oversight should be.

  4. However, the Institute is not a regulator. If Australia's AI Safety Institute identifies a capability that presents an unacceptable or inadequately controlled risk, which Australian authority has the ability to require action?

  5. For highly capable frontier systems, I believe this framework should include independent safety evaluation once defined capability thresholds are reached.

  6. If future AI systems develop increasingly sophisticated cyber capabilities, greater autonomy, the ability to conduct AI research themselves or the ability to circumvent human oversight, I do not believe voluntary commitments from the companies developing those systems will be sufficient.

Extracts from recommendations

Extracts chosen by an AI model, in the submitter’s own words. Listed in the same order as on its page in the interactive guide.

  1. I support the establishment of the Australian AI Safety Institute and the development of independent Australian expertise capable of evaluating advanced AI systems, understanding emerging risks and working with international AI safety organisations.

  2. I would like to see a proportionate regulatory framework specifically addressing frontier AI. This should not mean imposing the same regulatory burden on every AI application. Requirements should increase according to the capability of the system and the potential consequences associated with it.

  3. For highly capable frontier systems, I believe this framework should include independent safety evaluation once defined capability thresholds are reached. Areas warranting particular scrutiny could include sophisticated cyber capabilities, dangerous biological assistance, significant autonomous operation, autonomous AI research, evasion of human oversight and autonomous replication or persistence.

  4. For highly capable frontier systems, I believe this framework should include independent safety evaluation once defined capability thresholds are reached. Areas warranting particular scrutiny could include sophisticated cyber capabilities, dangerous biological assistance, significant autonomous operation, autonomous AI research, evasion of human oversight and autonomous replication or persistence. Where these thresholds are reached, progressively stronger safety requirements should apply.

  5. Ifa system demonstrates a genuinely dangerous capability and adequate safeguards cannot be demonstrated, there should be a mechanism allowing deployment or further development at that level to be temporarily paused while the risk is assessed.

  6. Frontier AI developers should also have obligations to report serious safety failures or unexpected dangerous capabilities to an appropriate independent authority.

  7. Employees raising credible concerns about significant safety risks should have appropriate whistleblower protections.

  8. Companies understandably need to protect intellectual property and commercially sensitive information. However, regulators also need sufficient access to independently assess the safety of systems with potentially significant consequences. We should not rely entirely on companies assessing and assuring governments of the safety of their own systems.

  9. Australia should continue working internationally towards shared safety evaluation standards, common approaches to dangerous capability thresholds and coordinated responses where significant risks emerge.

Read the original

Open in the interactive guide