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 535

Special Broadcasting Service (SBS)

Submitter group: Creative sector.

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.

SBS describes its AI framework and potential service and First Nations language benefits, alongside risks to news integrity and cultural material. Proposes risk-based regulation, transparency, attribution, copyright/ICIP protection, LLM inclusion in the News Bargaining Incentive and a digital competition regime, plus commensurate journalism support and investment in media/AI literacy. Says regulation should distinguish already-regulated media from AI companies.

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 it supports ruling out a text-and-data-mining exception, arguing such an exception would permit AI development using copyright material without permission, attribution or compensation.

    SBS welcomes the Government’s move to rule out a text and data mining exception in the Copyright Act. Such an exception would permit exploitation of copyright materials for AI model development without permission, attribution or compensation, in turn threatening the economic sustainability of Australia’s media, creative and cultural sectors.

  2. Argues that large language models should be brought within the News Bargaining Incentive framework.

    LLMs should be included in the News Bargaining Incentive framework because they are increasingly offering and being used to search for and access news, and offering content to consumers that purports to be news.

  3. Argues that AI systems should be required to disclose their data sources and decision-making methodologies.

    Guardrails or frameworks should require AI systems to disclose their data sources and decision making methodologies to address issues such as inaccuracies, misinformation, hallucinations and deepfakes, thereby helping media organisations and individuals identify reliable and trustworthy AI-generated or AI-assisted information or content.

  4. Says increasingly convincing AI-generated content raises concerns about deepfakes and manipulation, and that without human oversight and AI literacy these could affect trust in news and media.

    AI-generated content is becoming increasingly sophisticated and convincing, raising concerns in relation to deepfakes, manipulation and criminal activities. Without sufficient human oversight and public literacy, awareness, and education in relation to AI, these issues could impact trust in news and media overall.

  5. Argues that generative AI highlights the need for a flexible regulatory regime for digital platforms, administered by the ACCC.

    The rise of generative AI products and services and the rapid pace of change highlights the need for a flexible regulatory regime for digital platforms to address competitive and consumer risks, administered by the Australian Competition and Consumer Commission (ACCC).

  6. Argues that Australian law does not comprehensively protect Indigenous Cultural and Intellectual Property, creating a risk of its use by AI systems without consent.

    Australian law does not currently provide comprehensive protection for ICIP. In particular, existing intellectual property frameworks do not adequately protect cultural heritage that is intangible, collectively held and governed by customary laws, lore and protocols. This gap creates a risk that material embodying ICIP may be used by AI systems without consent, attribution, remuneration, or respect for cultural protocols.

Recommendations

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

  1. Match support for AI deployment with a commensurate increase in support for original, fact-based and independent public interest news and journalism.

    Support for AI deployment should be met with a commensurate increase in support for original, fact-based and independent public interest news and journalism which finds, verifies and ‘creates the record that AI draws from’

  2. Make significant investment in media and digital (AI) literacy to mitigate AI-based misinformation, scams and deepfakes, and measures to ensure Australians understand why and how to access reliable sources of information.

    there should be significant investment in media and digital (AI) literacy to mitigate the risk of AI-based misinformation, scams and deepfakes, as well as measures to ensure Australians understand why and how to access reliable sources of information.

  3. Make AI transparency, auditing and source attribution requirements an essential element of any policy or regulatory intervention.

    Requirements for AI transparency, auditing and source attribution must be an essential element of any effective policy or regulatory intervention

  4. Empower and resource regulators to investigate and enforce AI regulation.

    Regulators must be empowered and resourced to investigate and enforce regulation.

  5. Implement a flexible ACCC-administered digital competition regime; monitor AI self-preferencing, auto-renewal and tying, and level the playing field between big technology companies and Australian businesses.

    The rise of generative AI products and services and the rapid pace of change highlights the need for a flexible regulatory regime for digital platforms to address competitive and consumer risks, administered by the Australian Competition and Consumer Commission (ACCC).  SBS supports ongoing monitoring of anti-competitive behaviours and practices that could emerge in relation to AI, such as self-preferencing, auto-renewal and tying products and services, and supports recommendations to level the playing field between big tech companies and Australian businesses.viii

  6. Improve clarity on how existing competition and consumer law applies to AI-enabled goods and services.

    SBS would welcome improved clarity around the application of existing competition and consumer law to AI enabled goods and services.

  7. Keep government AI initiatives and funding ongoing and expand them to include support for ethical AI research, development of AI regulatory frameworks and public engagement on AI.

    It is essential that Government initiatives and funding are ongoing. Government initiatives should be expanded to include comprehensive support for research into ethical AI, the development of AI regulatory frameworks (or guardrails), and public engagement on AI matters.

  8. Consider ethical-AI certification for entities meeting industry standards or codes, taking account of Australia’s AI Ethics Principles, to help users identify compliant providers.

    Introduction of certification programs, when entities meet industry standards or codes of practice for ethical AI, taking into account Australia’s AI Ethics Principles, could help users, including SBS, identify AI providers which are in compliance with requirements.

  9. Consider supporting AI innovation hubs and incubators.

    The Government’s efforts could also include support for AI innovation hubs and incubators.

  10. Include LLMs in the News Bargaining Incentive framework.

    LLMs should be included in the News Bargaining Incentive framework because they are increasingly offering and being used to search for and access news

  11. Require through guardrails or frameworks that AI systems disclose their data sources and decision-making methodologies.

    Guardrails or frameworks should require AI systems to disclose their data sources and decision making methodologies to address issues such as inaccuracies, misinformation, hallucinations and deepfakes

  12. Explore transparency regulation of bots and crawlers to safeguard information integrity, counter disinformation and support copyright.

    Transparency regulation of bots and crawlers should be explored to safeguard information integrity, counter disinformation and support copyright.

  13. Require regular AI bias audits and correction, mandated detection/mitigation tools, diverse inclusive datasets, multidisciplinary teams and incentives for commercial training agreements on high-quality data.

    Regularly auditing AI systems for bias and implementing correction procedures when biases are detected are vital. Tools and methodologies are being developed, and should be mandated (e.g. via industry standards) to detect and mitigate bias in AI systems, and help to correct it. Mitigation work should involve using diverse and inclusive datasets and multidisciplinary teams, and incentives that encourage commercial agreements for AI training on high quality data.

  14. Adopt a risk-based approach to AI regulation focused on settings and activities posing the greatest potential for individual or societal harm, taking account of existing regulatory protections to avoid duplication or undue complexity, and adopting a consultative approach to avoid unintended consequences.

    SBS supports a risk-based approach to AI regulation, focusing on settings and activities that pose the greatest potential for individual and/or societal harm, taking account of existing regulatory protections to avoid duplication or undue complexity, and adopting a consultative approach to avoid unintended consequences.

  15. Have AI regulation differentiate between professional media organisations that are already highly regulated and AI companies.

    It is crucial that AI regulations differentiate between professional media organisations like SBS, which are already highly regulated and subject to rigorous standards of accuracy and accountability, noting AI companies are not subject to such requirements.

  16. Clarify and strengthen the copyright regime to recognise the value of copyright material for AI training, copyright owners' exclusive rights, and licensing by AI developers on fair terms.

    Australia’s copyright regime should be appropriately clarified and strengthened to recognise the value of access to and use of copyright material for AI model development and training, the exclusive rights of copyright owners to provide access to and use of copyright material for these purposes, and to ensure licensing of materials by AI developers takes place on fair terms.

  17. Explore regulation to enforce technical protection measures, without displacing broader legal protections and enforcement.

    Regulation to enforce technical protection measures should be explored, but should not displace broader legal protections and enforcement.

  18. Provide a sufficient policy or legislative response, including relevant guardrails, urgently on AI and copyright.

    sufficient policy or legislative response including relevant guardrails are urgently required in relation to AI and copyright.

  19. Develop standalone ICIP legislation through First Nations leadership, addressing traditional harms such as counterfeit art/merchandise and AI risks to First Nations knowledge and cultural expression.

    SBS welcomes the development of standalone ICIP legislation through a First Nations-led process and supports a framework that addresses both traditional harms, such as counterfeit art and merchandise, and emerging AI-related risks to the integrity and appropriate use of First Nations knowledges and cultural expressions.

  20. Address AI 'flooding' of the internet with unattributed information through industry-wide guardrails.

    AI ‘flooding’ the internet with information, including multiple instances of information without attribution to their sources, increases the difficulty of information verification, a process required to be performed by reputable news or media businesses, including SBS. This is also related to the issues of intellectual property outlined above, and should be addressed through industry-wide guardrails.

  21. Provide guidance on the impact of AI usage on SBS's sustainability.

    SBS welcomes guidance on the impact of AI usage on SBS sustainability.

  22. Champion SBS’s representative media to counter threats to democracy, cohesion and institutional trust.

    To counter these risks, representative media provided by SBS, which supports shared moments, a sense of belonging and social cohesion, should be championed.

Read the original

Open in the interactive guide