Submission 530
Alcohol and Drug Foundation (ADF)
Submitter group: Health.
An AI model identified the submitter group and wrote the summary, key points and recommendation summaries. Counts do not measure importance or merit.
Summary
Read the original submission for the full text.
The Alcohol and Drug Foundation, which operates the AI chatbot dib, describes opportunities and risks of AI for alcohol and other drug harm prevention. It describes bias, privacy and content moderation risks for people who use drugs, and asks governments to invest in trusted AI health information and set health-focused priorities in national strategies.
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.
Argues that general-purpose commercial AI tools are not designed to provide clinically governed alcohol and drug information or care.
However, general-purpose commercial AI tools are not designed specifically to provide clinically governed AOD information or care and cannot be relied upon to meet public-health objectives.
Describes its own chatbot dib as built around a safety-first model, with responses generated only from its curated, evidence-based content.
Unlike many publicly available AI systems, dib was designed from inception around a safety-first model. The platform utilises retrieval-augmented generation technology, meaning responses are generated only from ADF's evidence-based, human-curated content rather than information sourced from the open internet.
Says that in its survey of 3,638 Australians, helplines and professional support were rated most helpful, while AI tools such as ChatGPT were emerging as a preferred source of information.
In our survey of 3,638 Australians, 2 ADF found that, while dedicated helplines or online professional support are rated by respondents as the most helpful (69%) in relation to addressing AOD needs, AI tools, particularly ChatGPT are emerging as a preferred source for AOD information, valued for instant answers, ease of use, and availability.
Argues that AI has made it cheaper to create targeted alcohol marketing content, which it says evidence suggests strongly influences consumption.
AI has made it cheaper and easier to create large volumes of highly targeted alcohol marketing content, which the evidence suggests is highly influential on consumption patterns – particular for young people and people who drink in ways that pose a serious risk to their health.
Argues that Australian government health strategies show a lack of planning for safe, ethical and health-focused AI adoption.
While Australian governments are beginning to explore opportunities for digital healthcare in reducing AOD harms, there is a lack of planning for safe, ethical and health-focussed AI adaption.
Recommendations
Recommendations are shown beside the submission’s own words. Listed in the same order as on its page in the interactive guide.
Consider National Drug Strategy priorities: reliable AI AOD information/navigation, screening/referral/continuity, underserved specialist access, drug-harm detection and regulatory monitoring (including alcohol sales/marketing); prohibitions/heightened review, especially privacy/access decisions; and safe AI-supported AOD-care/data standards.
identification of shared priorities in the next iteration of the National Drug Strategy, which is currently under development. Considerations could include: • Identifying AI applications that governments should prioritise because they offer clear public-health benefits, such as: improving access to reliable, AI-generated AOD information and navigation; supporting screening, referral and continuity of care; extending access to specialist advice in underserved communities; identifying emerging patterns of drug-related harm; and regulatory data collection, monitoring and reporting (for example, of alcohol sales and marketing data). • AI uses that should be prohibited or subject to heightened scrutiny and review, particularly automated decisions affecting privacy, and access to health information. • Standards for safe AI-supported AOD care and data usage.
Invest in safe, trusted sources of AI-supported health information, to increase access, reduce stigma, provide anonymous and out-of-hours access, support regional and remote communities, complement existing services and engage cohorts facing barriers.
Governments should invest in safe, trusted sources of AI-supported health information to: • increase access to trusted health information • reduce stigma associated with seeking support • provide anonymous access to information and advice • support people in regional and remote communities • complement existing health services rather than replace them • provide support outside traditional service hours • engage cohorts that face barriers in accessing traditional health systems.
Have health strategies, including the next National Drug Strategy for AOD, set health-focused national AI priorities and support service-sector, lived-and-living-experience-guided research, evidence and product development for AI health information, care and support.
Recommendation 2: Australia’s health strategies – including, for AOD, the next National Drug Strategy –should identify priorities for governments in implementing national AI policies and services in a way that meets the health needs of Australians. This should include support for the service sector to undertake lived and living experience-guided research, evidence and product development to deliver AI-generated health information, care and support.