CMNS2530 Chap.10 Public Good, Social Change and Communicative AI
Public Good, Social Change and Communicative AI
Public-good communication raises questions about who can participate, whose concerns become visible and how platforms or organisations shape discussion. This chapter joins physical, digital and virtual public spheres with social change, unequal access, platform architecture and AI-mediated communication.
It evaluates automation through relationship consequences: disclosure, voice, trust, contestability, human review and organisational accountability remain essential even when software produces the response.
What this chapter covers
- 01
Physical, digital and virtual public spheres
- 02
Issues, interests and concerns as motivations for participation
- 03
Unequal access, privilege and strategic influence
- 04
Platform architecture and visibility
- 05
Data-based communication, bots and automation
- 06
Trust, disclosure, human review and organisational remedy
Worked example · free
Reopen a complaint bot's listening loop
- 2Explain how duplicate classification removes contextual stakeholder evidence.
- 2Connect generic closure to trust, voice and organisational listening.
- 2Preserve context, disclose automation and route vulnerability signals to accountable humans.
Key terms
- Public Sphere
- A space where people participate in discussion about shared issues, interests and concerns.
- Platform Architecture
- The design features that shape access, visibility, interaction and participation on a platform.
- Human Review
- Accountable examination by a person who can interpret context and change an automated outcome.
- Digital Divide
- Unequal access to devices, connectivity, skills, safety or meaningful digital participation.
- Contestability
- The ability to question, correct or appeal a consequential communication or decision.
- Communicative AI
- Automated systems that generate, mediate, classify or respond within organisational communication.
- Algorithmic Visibility
- The patterned amplification or suppression of content through platform or system processes.
- Public Good
- Collective benefit pursued through communication that supports access, participation and accountability.
Public Good, Social Change and Communicative AI FAQ
How does digital communication affect the public sphere?
Digital technologies can expand access, connection and participation, while platform design, safety, language, disability, time and algorithmic visibility still create inequalities. Analyse who can enter, speak, be heard and influence outcomes rather than treating online comments as a representative public by default.
What should be examined in AI-mediated communication?
Identify the data, function, disclosure, error consequence and route to accountable human review. Then assess effects on trust, organisation-public or employee relationships and listening. Technical accuracy matters, but so do access, explanation, contestability and remedy.
Are bots always harmful to relationships?
No. Bots can assist with scale, routine information and reporting. Harm arises when they conceal automation, remove context, simulate consensus, block human escalation or make consequential errors without remedy. Evaluate the function and safeguards in the specific organisational relationship.
What makes human review meaningful?
A reviewer needs relevant context, sufficient time, appropriate expertise and authority to change the outcome. A person who merely confirms a system decision does not provide effective contestability. Explain how stakeholders reach review, how urgent cases are prioritised, how corrections are communicated and how repeated errors change the wider system.
Does an online consultation represent the public?
Not automatically. Platform access, account requirements, moderation, visibility and safety shape who participates and whose contributions are seen. Identify excluded routes and missing groups, offer alternative access where possible and report how submissions affect decisions. A visible comment stream is evidence of participation by some people, not proof of representative consent.
How should AI communication be evaluated?
Measure accuracy and speed alongside understanding, access, voice, trust, resolution and remedy. Operational efficiency can conceal unresolved need if cases close quickly but stakeholders cannot correct errors or reach a person. Review difficult and unequal cases, disclose the automated role and preserve organisational responsibility even when an external vendor supplies the system.
Does a vendor carry all responsibility for an AI error?
No. A supplier may control part of the model or infrastructure, but the organisation selects the use, shapes integration and creates the stakeholder consequence. Contracts, audit access and incident reporting matter because the organisation must investigate, explain and provide remedy. If vendor limits prevent correction of a high-consequence output, the governance design is inadequate.
How should social-change claims be bounded?
Distinguish communication activity, participation, relationship change and broader social outcomes. Identify the period, affected groups, comparison and other contributing forces. A platform or campaign may make an issue more visible without redistributing influence or changing conditions. State the plausible contribution and the evidence needed before attributing complex change to one intervention.
Assessment move
For a public-good case, map access, voice, influence and response across physical, digital and virtual settings. Identify who cannot enter, whose contribution becomes visible and how input reaches a decision. For an AI case, trace data into classification or generation, then follow the automated action to disclosure, consequence, contestability, human review and organisational remedy.
Name any vendor boundary without transferring accountability away from the organisation. Compare efficiency measures with relationship outcomes and inspect difficult cases rather than only averages. Combine both maps when technology mediates participation.
Keep claims bounded because the acquired Week 11 and Week 12 material names some concepts without preserving every diagram or model detail, and do not invent a missing framework to create false certainty.
