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 484

Mr Edward Pierzchalski

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. Mitigating the risk of extinction from AI should be a global priority alongside other societal-scale risks such as pandemics and nuclear war.

  2. Frontier AI companies are explicitly trying to accelerate the development of frontier AI using frontier AI, with the goal of producing a feedback loop where the next generation of AI will mostly or entirely be built by the current generation of AI.

  3. This incident was far more serious than I expected, and far more serious than previous documented misalignment incidents.

  4. We should have well-resourced, independent reviews and reports for every potentially worrying incident, not just ones which happen to impact a third party or which OpenAI deems material.

  5. If we want to have any influence on the outcomes of future frontier AI, we need Australian interests to be represented and promoted by Australians actually, literally, physically talking to and working with the people building frontier AI, building working relationships and getting our hands dirty ensuring these models are safe.

  6. [But] each company—and country—is under intense competitive pressure not to unilaterally slow [the acceleration of AI capabilities].

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. We should have well-resourced, independent reviews and reports for every potentially worrying incident, not just ones which happen to impact a third party or which OpenAI deems material.

  2. Give Australia's AI Safety Institute (AISI) the resources required to at least match its sister organisations like the UK's AI Security Institute, or well-regarded independent organisations like METR.

  3. Get AISI access to do thorough safety and security evaluations of frontier models before deployment.

  4. Actively facilitate international coordination, as a diplomatic priority. - Work on getting the US and China in the same room to coordinate.

  5. Work on getting Anthropic, OpenAI, and the US government in the same room to coordinate.

  6. Work on turning "coordination" into "enforceable limits and restrictions on dangerous frontier AI development".

  7. we need Australian interests to be represented and promoted by Australians actually, literally, physically talking to and working with the people building frontier AI, building working relationships and getting our hands dirty ensuring these models are safe.

Read the original

Open in the interactive guide