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 269

International Centre for Missing & Exploited Children (ICMEC) Australia

Submitter group: Civil society.

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 that generative AI is accelerating child sexual exploitation and undermining detection, requiring urgent, concurrent responses. Recommends resourced child safety advice and cross-sector collaboration, near-term legislative reform, recurring research funding, and sovereign victim-identification tools trained on Australian data and governed under Australian law, paired with coordinated ongoing training.

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 synthetic material is making detection harder, leaving specialist victim identification and forensic teams overwhelmed.

    The result is a workforce overwhelmed: specialist victim identification and forensic teams, already managing significant caseloads, must now also determine what is ‘real’, synthetic, or deliberately altered - with offenders holding faster, more capable tools than those available to find and protect children.

  2. Describes research with ICMEC Australia and ACCCE, surveying nearly 1,900 Australians aged 16–18, finding AI involved in around one in four image-based child sexual abuse cases.

    New research from Adelaide University's Australian Cybercrime Observatory - conducted with ICMEC Australia and the Australian Federal Police's Australian Centre to Counter Child Exploitation (ACCCE), surveying nearly 1,900 Australians aged 16 to 18 - found AI is already involved in around one in four cases of image-based child sexual abuse.

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

  3. Recommends prioritising identified gaps in protections against harms to children for near-term legislation, ahead of broader AI regulatory reform.

    We urge the Committee to treat harms to children as requiring the most urgent regulatory response within its broader inquiry, with any gaps identified prioritised for near-term legislative action separate from, and ahead of, longer-term AI regulatory reform.

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

  4. Argues that AI companions are among the fastest-emerging risks to children online and that a single point-in-time response cannot track change on that timeline.

    Harms that were marginal two years ago, such as AI companions, now sit among the fastest-emerging risks to children online. A single point-in-time response cannot track change on that timeline.

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

  5. Argues that law enforcement AI investment should be matched with training, and workforce capability should not be resourced only after tools are deployed.

    Any investment in AI tools for law enforcement must be matched and coordinated with training for officers, and provision for ongoing training as tools and offending tactics evolve. Workforce capability should not be treated as a downstream item to be resourced after tools are deployed.

    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. Resource a dedicated child safety advisory mechanism in the AI Safety Institute and Office of AI to inform safety evaluations, standards and testing.

    The AI Safety Institute and Office of AI should include a dedicated child safety advisory mechanism, resourced to provide input into AI safety evaluation, standards and testing regimes.

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

  2. Support ongoing, resourced AI child-safety collaboration among government, law enforcement, industry and civil society, building on existing models such as ICMEC Australia's SaferAI for Children Coalition.

    ICMEC Australia calls for support of an ongoing, resourced mechanism for cross-sector collaboration between government, law enforcement, industry and civil society on AI and child safety, building on existing models such as ICMEC Australia's SaferAl for Children Coalition.

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

  3. Recommend near-term legislative action, ahead of broader AI regulatory reform, to close gaps in offence frameworks covering AI-generated CSAM and to set safety-by-design obligations for AI companion and chatbot products accessible to children.

    The Committee should recommend near-term legislative action - ahead of broader AI regulatory reform timelines - to close gaps in offence frameworks covering AI-generated CSAM at every point of its lifecycle, and to establish safety-by-design obligations for AI companion and chatbot products accessible to children.

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

  4. Review offences across the AI CSAM lifecycle: training and fine-tuning on abuse material, generating or attempting to generate synthetic or modified material, and storage, sharing and distribution.

    Existing offence frameworks should be reviewed to ensure they capture harm at every point of the AI CSAM lifecycle - from the use of child abuse material within datasets to train or fine-tune AI models, to the generation or attempt to generate synthetic or modified child abuse material, through to its storage, sharing and distribution.

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

  5. Consider obligations on developers and deployers of AI companion and chatbot products regarding child users, including safety-by-design requirements and age-appropriate safeguards.

    Consideration should also be given to obligations on developers and deployers of AI companion and chatbot products regarding child users - including safety-by-design requirements and age-appropriate safeguards.

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

  6. Establish obligations for platforms to identify young people's disclosures of sexual harm to AI chatbots and refer them to appropriate support.

    The same research found young people are already disclosing this harm directly to AI chatbots, with no consistent obligation on platforms to identify and refer these disclosures to appropriate support, reinforcing the need for the safety-by-design and referral obligations proposed above.

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

  7. Provide recurring research funding tracking young people's online behaviours, disclosure patterns and harms as AI evolves.

    The Committee should emphasise the need for ongoing funding for research tracking how young people's online behaviours, disclosure patterns, and experiences of harm are changing as AI becomes more embedded in their digital lives. This should be structured as a recurring commitment given how quickly the underlying technology and offending tactics are shifting.

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

  8. Scope and resource developing or procuring sovereign narrow AI, including facial recognition and image matching, for state/territory and federal victim-identification teams, trained on Australian data and governed under Australian law.

    The Government should scope and resource a program to develop or procure sovereign, narrow AI tools - including facial recognition and image-matching capability - for specialist victim identification teams within state/territory and federal law enforcement agencies, trained on Australian data and governed under Australian law.

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

  9. Prioritise the sovereign victim-identification AI program ahead of a longer-term AI capability roadmap.

    The Government should scope and resource a program to develop or procure sovereign, narrow AI tools - including facial recognition and image-matching capability - for specialist victim identification teams within state/territory and federal law enforcement agencies, trained on Australian data and governed under Australian law. We see a strong case for prioritising this ahead of a longer-term AI capability roadmap.

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

  10. Match any investment in AI tools for law enforcement with coordinated training for officers, including ongoing training as tools and offending tactics evolve.

    Any investment in AI tools for law enforcement must be matched and coordinated with training for officers, and provision for ongoing training as tools and offending tactics evolve.

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

Read the original

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

Open in the interactive guide