ABCDEFGHIJKLMNOPQRSTUVWX
1
IDAuthorCommentTheme
2
0Marjan Ehsassiis this complete? it seems to me that we don't want AI to replace learning or human judgementTheme 2: Human Oversight, Judgment, and the Limits of AI
3
1
Suzette Brooks Masters
I agree.Theme 2: Human Oversight, Judgment, and the Limits of AI
4
2
Suzette Brooks Masters
I also think it's true that more governments feel pressure to engage the public more meaningfully because of the growing lack of trust in those institutions and the threats to democratic legitimacy in the way it is currently practiced. The way it's worded now seems like an oversimplication. I think it's important to frame this in terms of the urgency to make the public feel like its voice is being heard, and to overcoming the barriers to making that a reality at scale.Theme 3: Trust Deficits — Toward Government and AI
5
3
Suzette Brooks Masters
But the truth is even without AI, these pressures to engage the public still exist.Theme 3: Trust Deficits — Toward Government and AI
6
4BonnieMight be helpful to more clearly articulate the targeted audience - eg. level of prior knowledge/experience expected of participants - it seems to move around a bit and I'm concerned that more experienced engagement practitioners looking for an AI overlay might get frustrated while engagement newbies will miss critical foundations.Theme 9: Course Structure, Scope, and Audience Calibration
7
5BonnieI wasn't sure where this case study fits - it is a demonstration that good community engagement does not always mean more engagement - particularly in communities who are fatigued or oversampled. And that AI enabled engagement doesn't have to mean faster or bigger - it can open up new opportunities for doing more with what you already have. Case study example: The City of Melbournes Knowledge Bank is community feedback analytics platform. Using custom-built NLP and machine learning, the Knowledge Bank makes it easy for council staff to understand community insights across topics, themes, and past engagement projects. This project shows leadership and excellence by addressing a longstanding sector-wide challenge: the difficulty of aggregating community engagement insights over time and across multiple consultations. While councils regularly consult on a wide range of topics, synthesising this feedback into actionable insights has traditionally been complex. The Knowledge Bank fills this gap by enabling staff to view all feedback in one place and using natural language processing to automatically categorise responses into themes such as transport, parks, and community safety. The team also chose to build sophisticated text analytics and machine learning capabilities in-house rather than outsourcing. This strategic decision ensures full accountability for the model, supports ongoing improvement, and builds valuable organisational expertise that can be applied to future data and innovation projects.Theme 9: Course Structure, Scope, and Audience Calibration
8
6BonnieThis project sets exemplary standards for data ethics and integrity. Automatic redaction protects community privacy by removing personal information. Feedback is presented in contributors’ original words, without editing or sanitisation, preserving authentic community voices. In addition, demographic benchmarking against ABS Census data provides transparent representation analysis, clearly highlighting any over- or under-representation across age, gender, and other diversity indicators. These practices ensure that decision-makers understand both the content and the context of community input.Theme 7: Data Protection, Privacy, and Compliance
9
7Bonniehttps://www.mav.asn.au/mavlab/futures/mavlab-innovation-library/knowledge-bank-working-groupTheme 9: Course Structure, Scope, and Audience Calibration
10
8BonnieCase Study = Advancing AI Innovation in Local Government - THe MAVlab team developed this in collaboration with 250 council staff, sector leaders and vendors in 2025. Includes a pretty comprehensive summary of the design research process and insights about AI use in councils - both in the statutory planning process and beyond. And an AI Use Case Library for Statutory Planning detailing 35+ use cases for AI integration and a series of inappropriate uses cases we recommend councils and vendors stay well away from. links below.Theme 9: Course Structure, Scope, and Audience Calibration
11
9Bonniehttps://www.mav.asn.au/mavlab/service-and-tech/aiTheme 9: Course Structure, Scope, and Audience Calibration
12
10Bonniehttps://www.mav.asn.au/__data/assets/pdf_file/0009/42993/MAVlab-AI-Planning-Report-Jul-2025.pdfTheme 4: Institutional Readiness and Resource Constraints
13
11Bonniehttps://www.mav.asn.au/__data/assets/pdf_file/0008/44738/Use-Case-Library-for-AI-in-Planning-for-Councils-Oct-2025.pdf
14
12BonnieThis work was in order to develop new procurement criteria to support local government source AI - details here: https://www.mav.asn.au/mavlab/service-and-tech/ai-procurementTheme 4: Institutional Readiness and Resource Constraints
15
13
Suzette Brooks Masters
and govt officials who can testify to how they overcame barriers and what this robust form of citizen engagement enabled.Theme 9: Course Structure, Scope, and Audience Calibration
16
14BonnieI would like to advocate for a deeper focus on risk and threat assessment and mitigation and management approaches. From my read, this set of modules assume that public participation in these processes will be with positive or benign intent. While this is always the hope, it is more common for public engagement processes - particularly technology mediated/AI enabled - to come under some form of attack - with the intent of either skewing results or disrupting the process. For all the reasons AI offers exciting and positive opportunities for enhanced community engagement, it also opens up negative and damaging pathways that need to be considered for the viability of the program and the safety/wellbeing of the participants. For consideration: Disruption from special interest groups. Special interest groups will target engagement processes anywhere in the world in order to forward their agendas. How will you ensure that the audiences you are seeking to work with/learn from are prioritised/not intimidated/overrun by outside actors? How will you manage/mitigate attack it your program/campaign is targeted? Coordinated information insertion attacks aimed at skewing results of engagement. On the more benign end, communities and/or special interest groups are increasingly literate in how to coordinate responses, providing talking points, letter/submission templates, coordinate response times for members etc to insert skewed data into processes. At the other end of the scale there are well funded coordinated bot swarms and targeted misinformation campaigns.Theme 6: Risk, Threat, and Adversarial Behavior
17
15BonnieMis and Disinformation - So much. All the time. On every channel. Before starting a process it is critical to assess what is likely to occur and how it will be moderated. Eg. what is the policy for inclusion/exclusion, what is the decision making/communication process, key messages/crisis response etc. Targeted FOI requests as mechanisms for information weaponisation - organisations must be considered/ undertake risk assessment on the types of information they are requesting as part of any engagement processes. Online surveys/open engagements publicly signal the types of data and information organisations hold. Consider how the information you are collecting can be weaponised. Abuse and Threats of Violence - Serious consideration must be given to the psychological and personal safety of government officials leading this work. Targeted abuse and threats of violence - through a variety of online and offline channels - are increasing for people who take publicly visible roles in community engagement. Particularly for women. Particularly for women of colour. What psychological and physical safety plans do you have in place? What is your process for threat escalation? What is your crisis response plan?Theme 6: Risk, Threat, and Adversarial Behavior
18
16Frenchen UppsalaI am missing the part about biased AI and therefor the possibility of biased suggestions and results. Is there a chance to incorporate it into the course ?Theme 2: Human Oversight, Judgment, and the Limits of AI
19
17
Suzette Brooks Masters
Is this the right title? It is perhaps civic or public engagement. Whether the inputs are used in a democratic way or in a way that strengthens democratic practice is a separate question. Perhaps that distinction needs to be made explicit and the engagement piece and accountability/strengthening democracy piece can be evaluated separately. They need not go together.Theme 9: Course Structure, Scope, and Audience Calibration
20
18Matt Ryanits not for this video (its actually mentioned in Module 1 but possibly needs be upfront and may only need to be on course website in written form), but somewhere perhaps something that gets expectations about the course in context. i.e. its short for a reason and part of that is its a complement not a replacement to public engagement education / training. In this course you'd learn about various points in an engagement project where integrating AI coulld assist.Theme 9: Course Structure, Scope, and Audience Calibration
21
19Javier PérezI think this Module Zero is the perfect place to introduce a word of caution on an issue that I believe InnovateUS agrees with. AI offers democracy new avenues, but not shortcuts. It is important to choose wisely in which areas or processes to introduce AI-enabled automation and in which it is preferable to focus on issues such as face-to-face deliberation, the ability to compromise and reach agreements, etc.Theme 2: Human Oversight, Judgment, and the Limits of AI
22
20Bonnie👍Theme 9: Course Structure, Scope, and Audience Calibration
23
21JOSE L. MARTÍThe title of the course is "democratic engagement". Does "democratic" adds anything to "public engagement" o simply "engagement"? If so, it should be clarified here in Module 0, I guess, or somewhere else. I can see a case to add "democratic" but should be clarifiedTheme 9: Course Structure, Scope, and Audience Calibration
24
22Ireland Twiggs_Marked as resolved_Theme 9: Course Structure, Scope, and Audience Calibration
25
23Ireland Twiggs_Re-opened_Theme 9: Course Structure, Scope, and Audience Calibration
26
24Marjan Ehsassidoes this assume that efficient and effective public conversations are impactful (in other words also hold gov't accountable)?Theme 1: Closing the Loop — Accountability and Impact
27
25faye liuThis is a great point, I think this could more clearly specify what is meant by efficient and effectiveTheme 9: Course Structure, Scope, and Audience Calibration
28
26faye liuI think the basis of 'platform-agnostic' is appropriate given the broad global audience, but my experience with supporting adults with learning AI more generally is that they are looking for more specific instruction for doing things - it might good to have one or two demo's included in the course? Or clearly flagged where they can watch someone doing the basicsTheme 9: Course Structure, Scope, and Audience Calibration
29
27Javier PérezSimilarly, access to some of the reference documents on guidelines for effective citizen participation could be facilitated. I'm thinking, for example, about the great 2022 OECD Guidelines for Citizen Participation Processes. https://www.oecd.org/en/publications/oecd-guidelines-for-citizen-participation-processes_f765caf6-en.htmlTheme 9: Course Structure, Scope, and Audience Calibration
30
28Ruthie Nachmanyshould we show some examples for how you can collect that engagement?Theme 9: Course Structure, Scope, and Audience Calibration
31
29Sarah HubbardMore context on Danielle's concept of the "spinal cord democracy": "Once these design principles were put to work in the constitution-writing of the late 18th through the 20th centuries, they yielded representative democracies, whose functioning depends entirely on what might be called the spinal cord of democracy. Every set of representative institutions depends on three core components: (1) channeling of citizen voice; (2) processing of public opinion by decision-makers to make policy decisions; and (3) enforcement of those decisions, delivery of government services, and judicial evaluation of them. The metaphor encapsulates all three of the foundational ideas outlined above. The popular basis of government and the reservation of power and ownership of state capacity to the people is conveyed through the importance of connecting the formal moments of government action (legislation, execution, and judicial evaluation) to the voices of the people." In our idea of the spinal cord - engagement is one of these key components of the spinal cord (voice). So here I wouldn't say "engagement as the spine" but maybe just "engagement as a core component" or something like that?Theme 9: Course Structure, Scope, and Audience Calibration
32
30Ricardo PoppiAgreed. Indeed, engagement without feedback leads to frustration :-(Theme 1: Closing the Loop — Accountability and Impact
33
31Stephen BuckleyHmmm .. Are there *other* core components that are distinctly separate from "Public Engagement"? (If not, then use "the" instead of "a".)Theme 9: Course Structure, Scope, and Audience Calibration
34
32Lex PaulsonI would suggest presenting these CI-related benefits in two categories: (1) how using CI raises the quality of decisions (cognitive diversity, idea synergy, 'many eyes' to find bugs) and (2) how using CI repairs broken civic relationships (knowledge and trust in public institutions, depolarization & bridging divides, growing leadership skills). Helene's new book is especially strong on #2.Theme 9: Course Structure, Scope, and Audience Calibration
35
33Mark WarrenAgreeTheme 9: Course Structure, Scope, and Audience Calibration
36
34Ireland TwiggsI am not sure if collective intelligence is the right word- maybe something like collective participation or collective lived experience?Theme 9: Course Structure, Scope, and Audience Calibration
37
35
Suzette Brooks Masters
this assumes that this is what government is doing. I don't think there is enough here about the threats governments at all levels are facing in a rising authoritarian environment. That needs to be a core part of the context setting.Theme 3: Trust Deficits — Toward Government and AI
38
36Hélène Ramarosonperhaps one could add something about the "double challenge" of trending trust issues/crises 1) towards governments/public institutions 2) towards AI --> how to tackle both...Theme 3: Trust Deficits — Toward Government and AI
39
37Mark WarrenI agreeTheme 3: Trust Deficits — Toward Government and AI
40
38Róbert BjarnasonTo be able to make the best possible public decisions for society in an ever more complex and fast moving world. (Public engagement is)Theme 9: Course Structure, Scope, and Audience Calibration
41
39William ZipseA public official is there to serve the public and not run them. -Gifford PinchotTheme 9: Course Structure, Scope, and Audience Calibration
42
40William ZipseI think it might be a good idea to outline that there are different objectives and expectations to engagement that change how public institutions engage the public (i.e. partnership, collaboration, comments, informing of decisions made). This may fit 2. e. in the outline as well.Theme 9: Course Structure, Scope, and Audience Calibration
43
41AnonymousAgree. We have also recently tried frame questions of what meaningful engagement means and to start with process and objectives, before jumping into technology. We tried to simplify different modes of engagement: static one-way, iterative one-way, to dynamic two-way loops. https://politicsir.cass.anu.edu.au/research/publications/civic-infrastructure-age-digital-engagementTheme 9: Course Structure, Scope, and Audience Calibration
44
42Aleks BerditchevskaiaAre we mostly considering GenAI? It might be obvious but still important to make that clear upfront. I think it's the type of AI that most public sector practitioners will be using (rather than predictive AI for e.g.). In our recent UK-wide public deliberations on AI we made the distinction between generative AI, predictive and perceptive AI. This seemed to resonate with people.Theme 9: Course Structure, Scope, and Audience Calibration
45
43Marjan Ehsassiand outline the potential risks involved with the use of AITheme 2: Human Oversight, Judgment, and the Limits of AI
46
44David FairmanCould add participation before deliberation. In terms of democratic principles, this is important in order for engagement not to be simply a data extraction exercise, but rather an experience that is interactive and to some degree empowering for citizens.Theme 9: Course Structure, Scope, and Audience Calibration
47
45Sarah HubbardWe could add in a fuller description of the "spinal cord of democracy" concept here? Will ask Danielle about this.Theme 9: Course Structure, Scope, and Audience Calibration
48
46
Suzette Brooks Masters
does it? what are the competing incentives that might prevent that from being true?Theme 1: Closing the Loop — Accountability and Impact
49
47Marjan Ehsassicommitment and accountabilityTheme 1: Closing the Loop — Accountability and Impact
50
48Kristin Jordahl HansenEven if public engagement doesn't generate better intelligence (and in some cases, it might not) it can generate more trust - which is also needed for governance to be effective. However, using AI can possibly reduce trust, so there is a tension.Theme 3: Trust Deficits — Toward Government and AI
51
49
Suzette Brooks Masters
AgreedTheme 3: Trust Deficits — Toward Government and AI
52
50Marjan Ehsassilearning and greater intelligenceTheme 9: Course Structure, Scope, and Audience Calibration
53
51Mario GrubisicA real-life example would be good to put here to make it more tangibleTheme 9: Course Structure, Scope, and Audience Calibration
54
52Mark WarrenI would add that public engagement can also bring people/organizations on board with decisions, generating not only democratic legitimacy, but also community capacities for addressing issues and problems. That is, while public engagement is mostly about voice and collective will-formation, it can result in expanded capacities for governance.Theme 4: Institutional Readiness and Resource Constraints
55
53David FairmanAgreedTheme 9: Course Structure, Scope, and Audience Calibration
56
54Antaraa VasudevAgreed - another positive outcome is better allocation of service delivery expenses and easier implementation of programs - both should be highlightedTheme 9: Course Structure, Scope, and Audience Calibration
57
55Javier PérezIn relation to the real-life example requested by Mario, I usually use the example of the Citizens' Assembly in Ireland, which, after years of political deadlock over how to incorporate the right to abortion into the Constitution, was able to provide political actors with a consensus solution that incorporated the values, principles, and concerns of citizens, paving the way for the institutionalization of the solution.Theme 9: Course Structure, Scope, and Audience Calibration
58
56Claudius Lievengain additional intelligence...Theme 9: Course Structure, Scope, and Audience Calibration
59
57Hélène RamarosonI would also add different perspectives and point of viewsTheme 9: Course Structure, Scope, and Audience Calibration
60
58Marjan Ehsassiadd collective intelligenceTheme 9: Course Structure, Scope, and Audience Calibration
61
59Lex PaulsonThis covers my #1 above; I would add something about #2, the benefits to civic relationships, trust and depolarization... this is likely to be front of mind for public servantsTheme 3: Trust Deficits — Toward Government and AI
62
60Ireland TwiggsIs it worth adding here as well as about lived experiences and varied perspectives and opinionsTheme 9: Course Structure, Scope, and Audience Calibration
63
61Ireland TwiggsI like the focus here on breadth and depth- important emphasis that traditional methods struggle to accomplish both of these at the same timeTheme 9: Course Structure, Scope, and Audience Calibration
64
62David FairmanThis would be a good place to reference IAP2 participation spectrum.Theme 9: Course Structure, Scope, and Audience Calibration
65
63JOSE L. MARTÍI'm missing more fine-grained discussion on the type of engagement that really serves the purposes of democratic engagement, because not any form of engagement worksTheme 9: Course Structure, Scope, and Audience Calibration
66
64Eve KleinThis mixes goals and methods - 'discovering what people care about' and 'closing the loop' are goals; 'facilitating deliberation' is a method. I think precision matters here since we're working to build skills around goal/method matching.Theme 9: Course Structure, Scope, and Audience Calibration
67
65Marjan Ehsassiconsensus buildingTheme 9: Course Structure, Scope, and Audience Calibration
68
66Alisson BrunoIn some cases, public participation does not directly influence final decisions. For example, when citizens submit questions during public hearings and receive answers, the feedback loop may be considered closed, even if their input does not lead to concrete changes in policy. This suggests that “closing the loop” does not necessarily imply direct impact on decision-making, but can also refer to responsiveness and transparency.Theme 1: Closing the Loop — Accountability and Impact
69
67William ZipseI would also add the other end of the spectrum that public participation could also be used to solicit ideas from the public.Theme 9: Course Structure, Scope, and Audience Calibration
70
68AnonymousHere is one typology of participation that might be helpful: https://engagementinstitute.org.au/resources/Theme 9: Course Structure, Scope, and Audience Calibration
71
69Claudius Lievenhow citizens input contributed to shaping decisions --> adjust expectation levelTheme 1: Closing the Loop — Accountability and Impact
72
70Róbert BjarnasonThe possibility of highly customized one-off engagement software solutions built on-demand by the project team - there are always compromises with general solutions, where we are with AI today, custom solutions for each task is already a reality for coder, later this year likely for everybody.Theme 9: Course Structure, Scope, and Audience Calibration
73
71Dino CantúI would add one more bullet on "risks and limitations" warning stories, "when AI goes wrong and how to prevent it"Theme 2: Human Oversight, Judgment, and the Limits of AI
74
72Lex PaulsonI would add in the use of AI to help ideation & learning, e.g. helping citizens flesh out their ideas or use analogies from other fields. Mark Klein and I are working on this at UM6P, see also Ethan Mollick: https://www.oneusefulthing.org/p/against-brain-damageTheme 2: Human Oversight, Judgment, and the Limits of AI
75
73Claudius LievenAI can explain facts, improve presentation and summarize feedback...Theme 2: Human Oversight, Judgment, and the Limits of AI
76
74Bonniecome backTheme 9: Course Structure, Scope, and Audience Calibration
77
75Frenchen UppsalaWill you be addressing the factor "data location" for different AI usages? Uploading data to train or use AI has the difficulty of choosing very specifically and with expertise which "product" to useTheme 7: Data Protection, Privacy, and Compliance
78
76Hélène Ramarosonone can add the stage of selection/stratification of participants (i.e. design of participatory framework)Theme 8: Participant Selection — Language, Methods, and Representativeness
79
77Javier PérezAgree! I think AI can help public institutions overcome the bubble of the usual suspects and incorporate a greater diversity of voices into participatory processes, including actors who have at some point connected with that administration or who are simply active and present in the public debate.Theme 5: Equity, Inclusion, and Reaching Beyond Usual Suspects
80
78Claudius LievenTypically, public input is collected during the participation process. At this point, the information base required prior to the process is addressed. This should be differentiated.Theme 9: Course Structure, Scope, and Audience Calibration
81
79Róbert BjarnasonMaybe "public input" is too narrow as "Deep Research" on the web, on gov data etc could be useful.Theme 9: Course Structure, Scope, and Audience Calibration
82
80
Suzette Brooks Masters
True but at a more basic level it may just uncover broad based agreement that the polarizing political environment prevents govt from acting upon.Theme 3: Trust Deficits — Toward Government and AI
83
81Antaraa VasudevAI can also assist with assessing the quality of policy documents across parametres such as: - Comprehension - Justification - Comprehensiveness - Analysis of proposed benefits and costs Before these documents are made public. Eventually, citizen FAQs can also be generated from existing data setsTheme 2: Human Oversight, Judgment, and the Limits of AI
84
82Antaraa VasudevThe rubric of assessment is recommended to be designed by qualified researchers and practionersTheme 2: Human Oversight, Judgment, and the Limits of AI
85
83Ricardo PoppiAI can help on powering conversational interfaces for participation, specially using voice input, required for low literacy contexts. It lowers language barriers and lower the cognitive load of the participation processTheme 5: Equity, Inclusion, and Reaching Beyond Usual Suspects
86
84Mario GrubisicIt can go even further when it comes to personalizing messages (e.g. through the use of AI-generated avatars, such as Synthesia), which is especially useful in the outreach phase. This is even more relevant in cases where the target group is more removed from democratic decision-making, e.g. rural areas, minorities.Theme 5: Equity, Inclusion, and Reaching Beyond Usual Suspects
87
85Lex Paulson'explain processes and help citizens develop their ideas'Theme 5: Equity, Inclusion, and Reaching Beyond Usual Suspects
88
86Antaraa VasudevAgreed - also provide clarifications in plain languageTheme 5: Equity, Inclusion, and Reaching Beyond Usual Suspects
89
87
Suzette Brooks Masters
what about digital divide issues?Theme 5: Equity, Inclusion, and Reaching Beyond Usual Suspects
90
88JOSE L. MARTÍThe Habermas Machine guys are working now on the version 2.0 which they claim will be able to predict deliberative results in a particular groupTheme 2: Human Oversight, Judgment, and the Limits of AI
91
89Mark WarrenI would add that, depending on how organizers wish to scale to broader publics, it might be useful to include AI-driven deliberative platforms that can scale, such as the one developed by Stanford's Deliberative Democracy Lab. Often public entities are interested not just in gaining informed and representative input, but also reaching as many people as possible.Theme 8: Participant Selection — Language, Methods, and Representativeness
92
90JOSE L. MARTÍThat's crucial, I agree!Theme 9: Course Structure, Scope, and Audience Calibration
93
91Mark WarrenI would also note that few AIs are good at providing justifications for decisions (depending upon what is asked). Not an expert here, but apparently the newer versions of Claude are superior in this dimension to ChatGPT and others.Theme 2: Human Oversight, Judgment, and the Limits of AI
94
92Ireland TwiggsWhat exactly is considered “toxic” behavior?Theme 2: Human Oversight, Judgment, and the Limits of AI
95
93Antaraa VasudevAI can also support in overcoming language barriers, through real time translation and transcriptionTheme 5: Equity, Inclusion, and Reaching Beyond Usual Suspects
96
94Marjan EhsassiI am be the only here but the use of absolute verbs seems misplaced ... the use of AI "can make" etcTheme 9: Course Structure, Scope, and Audience Calibration
97
95Bonniestrong agreeTheme 9: Course Structure, Scope, and Audience Calibration
98
96Antaraa VasudevAn example here is the Government of Maharashtra - which engaged 0.4 million citizens across 37 districts of their State, through a chatbot to share their vision for development. 1 in 11 citizens responded via voice note, with feedback coming in across 3 Indic languages through handwritten letters, voice notes, chat responses and detailed documents. The Government was able to synthesise and publish all feedback, and it yielded a detailed vision document which has an anticipated impact across 22 policy areas in the State. More here: https://medium.com/civis-vote/when-citizens-speak-policy-listens-9c0f13056004Theme 5: Equity, Inclusion, and Reaching Beyond Usual Suspects
99
97Marjan Ehsassiresponsiveness is not the same as accountabilityTheme 1: Closing the Loop — Accountability and Impact
100
98Antaraa VasudevAI can also review public input and categorise it differentiating between degrees of public spirited input recievedTheme 2: Human Oversight, Judgment, and the Limits of AI