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 169

QUT Digital Media Research Centre

Submitter group: Universities.

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.

The QUT Digital Media Research Centre recommends funding humanities, arts and social sciences in sovereign AI, a digital duty of care, inclusive development, language datasets and media literacy. It describes abuse and misinformation risks, while recognising beneficial accessibility and advocacy uses and warning that catch-all regulation can harm disempowered groups and technical safeguards have uneven reliability.

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 Australia's humanities and social science research on AI should be recognised as a core part of the national AI research base.

    Australia is home to a vibrant and internationally leading humanities and social science research ecology working on artificial intelligence, with expertise spanning media and communication, digital society, human-computer interaction, sociology, anthropology, law, ethics, and cultural studies. These capabilities must be recognised as a core part of Australia’s strategic AI research base.

    A program corrected obvious text-recognition errors in this quote for display.

  2. Argues that AI image and video generation can be used to produce child sexual abuse material and non-consensual sexual content.

    Of particular and urgent concern is that AI-powered image and video generation technologies can be utilised to produce Child Sexual Abuse Material (CSAM) as well as other kinds of non-consensual sexual content which has, to date, overwhelmingly targeted young women.

    A program corrected obvious text-recognition errors in this quote for display.

  3. Says its findings indicated that, while there were some guards against conspiracy theories being promoted by generative AI chatbots, they were applied in a very selective manner.

    Our findings indicated that while there were some guards against conspiracy theories being promoted by GenAl chatbots, they were applied in a very selective manner.

  4. Says its forthcoming research found chatbot users writing in languages other than English are more likely to encounter conspiratorial content.

    This is evidenced in our forthcoming paper, demonstrating that chatbots users who utilise languages outside of English are more likely to encounter problematic, conspiratorial content from generative AI chatbots.

    A program corrected obvious text-recognition errors in this quote for display.

  5. Argues technical safeguard requirements should reflect current evidence of tools' reliability, avoiding greater confidence in technical solutions than the evidence supports.

    Requirements around technical safeguards should therefore be informed by a clear and up-to-date understanding of what these tools can reliably achieve. Without a critical engagement with technical effectiveness claims and the evidence underpinning these claims, there is a risk of placing greater confidence in technical solutions than the available evidence supports.

  6. Says QUT research finds judgements about AI-generated content culturally specific and politically contested, and that catch-all regulation can harm disempowered groups.

    QUT researchers have shown that judgements about the value and authenticity of AIGC are culturally specific and politically contested.2” These differences mean that catch-all approaches to regulating AI-generated content can be harmful and punitive to disempowered groups.

    A program corrected obvious text-recognition errors in this quote for display.

Recommendations

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

  1. Draw on and fund HASS research on AI alongside interdisciplinary efforts, so HASS disciplines participate in developing sovereign AI capabilities.

    It is vital that HASS disciplines participate in developing AI sovereign capabilities in Australia’s national interest. This means drawing on and funding HASS research on AI alongside interdisciplinary efforts to develop, study, and understand the implications of AI on Australian society.

    A program corrected obvious text-recognition errors in this quote for display.

  2. Provide significant investment in talent, training, research funding and infrastructure support to embed HASS capabilities into the sovereign AI effort.

    Embedding HASS capabilities into the sovereign AI effort requires significant investment in talent and training, as well as research funding and infrastructure support.

    A program corrected obvious text-recognition errors in this quote for display.

  3. Reverse policy settings, including Job-ready Graduates funding arrangements, that it says weaken the humanities and social science pipeline.

    This in turn will entail the reversal of policy settings that have weakened the HASS pipeline, including the Job-ready Graduates funding arrangements, which impose substantially higher student contributions on many humanities and social science disciplines.

  4. Recognise in law that AI companies share with platforms a direct duty of care to prevent harms such as gender-based violence and child exploitation.

    We recommend the recognition in law that AI companies share with platforms a direct duty of care to prevent their tools from being used to perpetrate gender-based violence, child exploitation, and other harms.

    A program corrected obvious text-recognition errors in this quote for display.

  5. Provide dedicated systems and support for an inclusive approach to the design, development and deployment of AI in Australia, so benefits and harms are not inequitably distributed.

    We recommend that dedicated systems and support be provided for an inclusive approach to the design, development and deployment of AI in Australia, to ensure the benefits and harms are not inequitably distributed

    A program corrected obvious text-recognition errors in this quote for display.

  6. Firmly position AI systems and their developers within the purview of the Digital Duty of Care, to enable strict and enforceable laws against malicious AI-generated audiovisual content.

    We recommend that the Committee firmly positions AI systems and their developers within the purview of the DDOC; doing so will empower legislators to develop strict and enforceable laws to protect Australians of all ages and backgrounds from the risk of malicious AI-generated audiovisual content.

    A program corrected obvious text-recognition errors in this quote for display.

  7. Develop comprehensive legislation to mitigate the risk of AI-assisted exploitation.

    It is critical that Australia develop comprehensive legislation to mitigate the risk of AI-assisted exploitation.

    A program corrected obvious text-recognition errors in this quote for display.

  8. Invest in low-resource and digitally disadvantaged language datasets to inform chatbots and protect culturally and linguistically diverse communities from problematic output.

    It is the responsibility of the Australian Government to ensure investment in low-resource and digitally disadvantaged language datasets to inform chatbots and protect culturally and linguistically diverse communities from problematic output

  9. Place pressure on AI companies to implement features reinforcing reminders that GenAI chatbot output is flawed.

    place pressure on AI companies to implement features to reinforce reminders that GenAl chatbot output is flawed

    A program corrected obvious text-recognition errors in this quote for display.

  10. Establish an advisory body to promote safer chatbot experiences.

    establish an advisory body to promote safer chatbot experiences.

  11. Include educational resources in AI policy so Australians can recognise AI-generated media content and its possible impact.

    We recommend that Australia’s policy direction must include educational resources developed to ensure that Australians are equipped to recognise and understand what AI-generated media content is and how it might impact them.

    A program corrected obvious text-recognition errors in this quote for display.

  12. Ensure tools for verifying AI-generated content are easily accessible, so the burden of proof does not fall to consumers.

    while it is important to develop technical methods of verifying AI-generated content, it is also necessary to ensure that such tools are easily accessible so that the burden of proof does not fall to consumers.

    A program corrected obvious text-recognition errors in this quote for display.

  13. Inform requirements around technical safeguards by a clear and up-to-date understanding of what these tools can reliably achieve.

    Requirements around technical safeguards should therefore be informed by a clear and up-to-date understanding of what these tools can reliably achieve.

  14. Invest in AI-content literacy programs and demand disclosure from media outlets and platforms, using an approach that learns from and uplifts diverse existing community capabilities.

    It is important that we continue to invest in programs that can uplift people’s ability to discern and evaluate AI-generated content, in addition to demanding more truth and transparency from media outlets and platforms (e.g., disclosure mechanisms such as labels). This should involve an approach to AI literacy that both learns from and uplifts the already existing and diverse community capabilities.

    A program corrected obvious text-recognition errors in this quote for display.

  15. Have transparency regulation go beyond binary classification of content as AI-generated or not, toward detailed, context-aware information about how and why GenAI was used.

    regulatory efforts around transparency should go beyond binary classifications of content as being AI-generated (and therefore ‘fake’) or not AI-generated (and therefore ‘real’), Instead, detailed, context-aware information about how and why GenAl was used in specific cases would enable audiences to discern whether GenAl use is appropriate and responsible.

    A program corrected obvious text-recognition errors in this quote for display.

  16. Ensure and support digital accessibility, including AI-related tools, for equal participation in social interactions.

    It is vital for people’s equal participation in social interactions that digital accessibility is ensured and supported, including the use of AI-related tools.

    A program corrected obvious text-recognition errors in this quote for display.

  17. Improve public understanding of platform governance so media industries and audiences understand how AI-generated content is moderated, verified, contested, amplified and transformed.

    Improved public understanding of platform governance is urgently needed to ensure that media industries and their audiences better understand how AIGC is moderated, verified, contested, amplified and transformed as it circulates.

Read the original

In some quotes a program corrected obvious text-recognition errors for display.

Open in the interactive guide