Require meaningful human oversight of consequential AI-assisted decisions
37 submissions. Each submission counts once for this shared ask. Counts do not measure importance or merit.
A shared ask groups requests for substantially the same action from three or more submissions when the group was created. Groups remain listed if withdrawals reduce their counts. AI models summarised recommendations, grouped them and checked the groups. The label summarises the requests; it is not a quote. A shared ask does not mean agreement on anything else.
Theme: Human oversight and accountability
Recommendations from submissions
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. Extracts are chosen by an AI model. Submissions are listed in submission number order; recommendations follow their order on each submission’s page.
OpenAI Submission 11
Require national security uses of AI to operate within Australian law, keep meaningful human judgement over consequential decisions and be subject to democratic oversight by institutions with technical expertise and appropriate access.
National security uses must operate within Australian law, preserve meaningful human judgement over consequential decisions, and be subject to effective democratic oversight. The institutions responsible for that oversight should have the technical expertise and appropriate access needed to scrutinise how AI contributes to government decisions.
Centre for Public Integrity Submission 14
Ensure that where legislation requires human judgment or discretion, the responsible decision-maker genuinely exercises it rather than endorsing an automated outcome.
Where legislation requires the exercise of human judgment or discretion, administrative arrangements must ensure that the responsible decision-maker genuinely exercises that power rather than simply endorsing an automated outcome.
Shop, Distributive and Allied Employees' Association (SDA) Submission 17
Where automated systems contribute to discipline or performance outcomes, disclose their standards to workers in advance and require meaningful human assessment of individual circumstances before disciplinary consequences.
Where an automated system contributes to a disciplinary or performance outcome, the standard it applies must be disclosed to the worker in advance, and a human decision-maker must meaningfully assess individual circumstances before any disciplinary consequence follows.
Environment Institute of Australia and New Zealand (EIANZ) Submission 31
Keep final statutory rejections, regulatory assessments and assurance conclusions subject to human judgement, appropriately qualified professional oversight and explicit statutory accountability.
final statutory rejections, regulatory assessments and assurance conclusions should remain subject to human judgement, appropriately qualified professional oversight and explicit statutory accountability.
Suicide Prevention Australia Submission 36
Require meaningful human oversight, independent evaluation before and during AI deployment for accuracy, bias and unintended harms, ongoing monitoring, and clear accountability for AI-informed decisions.
The Australian Government should require meaningful human oversight and independent evaluation of AI systems before and during deployment, including assessment of accuracy, bias and unintended harms, with ongoing monitoring and clear accountability for decisions informed by AI.
ADEIA, UNSW and the ARC Centre for Next-Gen Architectural Manufacturing Submission 53
Make supported domain systems professional decision-support rather than automated certification, retaining identifiable current authoritative sources and human professional responsibility.
These systems should be designed as professional decision-support, not automated certification. Authoritative sources must remain identifiable and current; human professionals must retain responsibility;
Institute for Integrated Economic Research - Australia Submission 65
Have the Australian Standards for AI incorporate cognitive impact and define meaningful human oversight operationally; make the automation setting for each significant public-sector decision function a documented, disclosed, reviewable decision rather than a vendor default.
That the Australian Standards for AI incorporate cognitive impact as a design consideration and define meaningful human oversight operationally, specifying the conditions under which a person is positioned to interrogate, override and take responsibility for a system’s output; and that the automation setting for each significant public-sector decision function be a documented, disclosed and reviewable decision rather than a default inherited from the vendor.
Western Sydney Regional Organisation of Councils (WSROC) Submission 87
Recommendation 3: national AI policy should set clear expectations that public-sector AI systems have proportionate human oversight, accountability and avenues for review, particularly where AI may influence decisions affecting individuals or communities.
Recommendation 3: National AI policy should establish clear expectations that public-sector AI systems must have proportionate human oversight, accountability and avenues for review, particularly where AI may influence decisions affecting individuals or communities.
A program corrected obvious text-recognition errors in this quote for display.
Centre for Law as Protection, Deakin University Submission 98
Legislate a binding 'meaningful human control' standard specifying minimum requirements, such as reviewer access to underlying evidence, adequate time and genuine authority to override.
Legislate a binding ‘meaningful human control’ standard. The standard should specify minimum requirements for what counts as ‘meaningful’ — e.g. the human reviewer must have access to underlying evidence (not just a confidence score), adequate time to review, and genuine authority to override.
Soroptimist International Australia Inc Submission 115
Require meaningful human oversight of consequential decisions, with a clear right to explanation, review and appeal.
Requiring meaningful human oversight for consequential decisions, with a clear right to explanation, review and appeal.
Professor Alysia Blackham Submission 124
Require oversight of automated decisions or outcomes by a human decision-maker ('human-in-command').
Oversight of automated decisions or outcomes by a human decision-maker – an approach described as ‘human-in-command’.
Nuvento Submission 128
Give authorised staff the evidence, time, training and authority to reject an AI output, and do not make a model score or generated allegation the sole basis for adverse action.
Meaningful human review: give authorised staff the evidence, time, training and authority to reject an output. A model score or generated allegation should not be the sole basis for adverse action.
Infosys Submission 141
Keep consequential decisions accountable to identified people, supported by auditable decision records, and develop and resource meaningful human oversight as a skill.
Infosys recommends that accountability for consequential decisions remain with identified people, supported by the auditable decision records described elsewhere in this submission, and that the capability to exercise meaningful oversight be treated as a skill to be developed and resourced in its own right.
United Nations Special Rapporteur on human rights and counter-terrorism Submission 160
Require AI to support human decisions, with effective qualified oversight, heightened scrutiny of elevated rights risks, and prevention and minimisation of unintended or unanticipated human-rights harms.
AI should support, not replace, human decision-making in counter-terrorism contexts. Qualified personnel must retain effective oversight of operational uses of AI, with heightened scrutiny where impact assessments have identified elevated risks to human rights. Human oversight should aim to prevent and minimise risks to human rights from the operation of AI systems, including from unintended or unanticipated consequences.
SA Unions Submission 175
Prohibit employers from using AI to make decisions affecting workers without human oversight.
And employers should be prohibited from using AI to make decisions affecting workers without "human oversight".
National Farmers' Federation Submission 188
Make consequential decisions explainable, reviewable and capable of appropriate human intervention.
Contestability and human oversight: consequential decisions should be explainable, reviewable and capable of appropriate human intervention.
AQ Intelligence (AQI), Guardrail Technologies Australia (GTA) and First Nations Strategic Alliance (FNSA) Submission 202
Preserve meaningful human oversight roles in high-stakes settings rather than allow full automation.
In high-stakes settings, policy should preserve meaningful human oversight roles rather than allow full automation — protecting both safety and skilled employment.
Centre for Future Work Submission 221
Require meaningful, direct human intervention in critical employment-related decisions.
Critical employment-related decisions must not be made without meaningful and direct human intervention.
Samantha Connor AM Submission 229
Guarantee meaningful human determination before automated/AI-assisted refusal, termination, suspension or material reduction of essential government payments, disability supports, aged-care entitlements, housing services or comparable benefits.
Establish an enforceable right to meaningful human determination before an automated or AI-assisted system can refuse, terminate, suspend or materially reduce an essential government payment, disability support, aged care entitlement, housing service or comparable benefit.
Cordexa Submission 230
Require an evidence record and human approval gate in Commonwealth AI procurement affecting people or public decisions.
Use Commonwealth procurement to set the floor: require the evidence record and a human approval gate in every AI procurement that affects a person or a public decision.
Deputy Submission 237
Apply controls proportionately to task consequences: strong access controls, logging, testing, escalation pathways and mandatory human review for AI materially affecting employment, pay or safety.
Apply risk controls proportionately to the consequence of the task: strong access controls, logging, testing, escalation pathways and mandatory human review for any AI materially affecting a person's employment, pay or safety.
Dr Fan Yang and Dr Robbie Fordyce Submission 244
Require workers to communicate with, and be assessed by, human decision-makers.
Workers must always communicate with, and be assessed by, human decision-makers in a humanity centric workplace: Workers experience physical and mental strain from working with or for AI, alongside weakened workplace bargaining power.
National Employment Services Association (NESA) Submission 267
Require program Deeds and guidelines to ensure meaningful human oversight and named accountability for decisions affecting eligibility, payments, compliance, service allocation or other rights and interests, equally for departmental and provider systems.
The Government should require, in program Deeds and guidelines, meaningful human oversight and named accountability for any decision affecting eligibility, payments, compliance, service allocation or other rights and interests. This should apply equally to departmental and provider systems and responsibilities.
Office of the Special Envoy to Combat Islamophobia (OSECI) Submission 287
Keep AI systems in law enforcement, immigration, counter-terrorism and national security subject to meaningful human oversight and annual independent auditing, so assessments are transparent, accountable and do not produce discriminatory outcomes for religious, ethnic or cultural communities.
Ensure AI systems deployed in law enforcement, immigration, counter-terrorism and national security remain subject to meaningful human oversight and annual independent auditing, so that automated assessments are transparent, accountable and do not produce discriminatory outcomes for religious, ethnic or cultural communities.
Australian Lawyers Alliance (ALA) Submission 310
Require adverse determinations on liability, eligibility, treatment, benefits, impairment or quantum to be made by a human decision-maker who has considered individual circumstances and can depart from the AI output.
Any adverse determination concerning liability, eligibility, treatment, benefits, impairment or quantum should be made by a human decision-maker who has considered the person’s individual circumstances and has authority to depart from the AI-generated output.
Finance Sector Union Submission 332
Require an accountable human to make every employment decision, including hiring, performance appraisal and management, redundancy and dismissal, and require employers to tell workers when AI has informed a decision about them.
Require an accountable human to make every employment decision, including hiring, performance appraisal and management, redundancy and dismissal, and require employers to tell workers when AI has informed a decision about them.
Mr Philip Gluyas Submission 366
Using ADM technology is useful, but it must be backed up by human oversight to ensure accuracy. This will provide confidence in it's use both by the departments and by the community that rely on said departments for the proper delivery of government services.
Ms Alana Kane Submission 392
For high-risk applications, require organisations to demonstrate accountability, permitted AI decisions, human-approval points, stopping and access-revocation mechanisms, decision auditing and post-deployment monitoring.
For high-risk applications, organisations should be required to demonstrate: • who is accountable for the system; • what decisions the AI is permitted to make; • when human approval is required; • how the system can be stopped; • how access can be revoked; • how decisions can be audited; • and how the system is monitored after deployment.
Mr Craig Thomler Submission 466
Require accountable human decision-making where Commonwealth AI materially affects rights, services, eligibility, employment, procurement, compliance or enforcement.
Require accountable human decision-making where Commonwealth AI materially affects rights, services, eligibility, employment, procurement, compliance or enforcement.
Ms Serena Rix Tripathee Submission 478
Require meaningful human oversight, with appropriately qualified humans able to review and override significant automated decisions.
Require meaningful human oversight and the ability for appropriately qualified humans to review and override significant automated decisions.
Ms Julie Green Submission 487
Require human decision-makers for AI-informed decisions affecting children or vulnerable people; AI remains a decision-aid only. Decision-makers must be able to see and challenge its reasoning; families must be informed and able to contest use.
Require a human decision-maker in the loop for any AI-informed decision affecting a child or vulnerable person. AI should never be the decision-maker, at best a decision-aid. That person must be able to see and challenge the tool's reasoning, and the family must be told a tool was used and given a genuine way to contest it.
Australian Nursing & Midwifery Federation (ANMF) Submission 522
Do not use AI to analyse performance, discipline, recruit, roster or allocate work without transparency, meaningful human review and a clear way for workers to challenge a decision.
AI should not be used to analyse performance, discipline, recruit, roster or allocate work without transparency, meaningful human review and a clear way for affected workers to challenge a decision.
Health Services Union (HSU) Submission 529
Mandate meaningful human-led clinical oversight and input into AI system decisions, particularly those affecting vulnerable groups, funding allocations and access to care.
Mandate meaningful human-led clinical oversight and input of AI system decisions particularly, where they affect vulnerable groups, funding allocations and access to care.
Community and Public Sector Union (CPSU) Submission 534
Keep human decision-makers responsible and accountable for government decisions, with informed human oversight embedded in all decision-making processes.
Human decision-makers must remain responsible and accountable for government decisions. Where AI is used to support decision-making processes, there is a risk of automation bias, whereby decision-makers may defer uncritically to AI-generated outputs. To mitigate this risk, informed human oversight must be embedded in all government decision-making processes.
Professor Victor Minichiello Submission 538
Address values, accountability and purposes in AI design; require meaningful human oversight of consequential systems, with responsibility for risk review and intervention and preservation of human accountability.
Public-interest AI policy should therefore address not only identifiable harms, but also the values, accountability arrangements and purposes built into systems from the beginning (Floridi et al., 2018). For consequential AI systems, this should include meaningful human oversight, with clear responsibility for reviewing significant risks, intervening where necessary and ensuring that automated systems do not displace human accountability.
Ms Jessica Browne Submission 555
Where human oversight is legally or professionally required, provide reasonably necessary information, authority, evidence and controls; do not rely on “human in the loop” when technology removes visibility or control.
Where law, regulation or professional standards require human oversight, the accountable person must have the information, authority, evidence and controls reasonably necessary to perform that oversight. "Human in the loop" should not be relied upon where technology progressively removes the human's visibility or control.
Ms Jodie-Ann Jones Submission 557
Require a contestability and human-oversight standard for automated decision-making in emotionally or psychologically sensitive AI tools, distinct from general AI ethics guidance.
Require a contestability and human-oversight standard for automated decision-making in emotionally or psychologically sensitive AI tools specifically, distinct from general AI ethics guidance, given how different the stakes are when the "decision" is whether a human being is told someone may be at risk.