An exploratory review of ChatGPT, Gemini and Claude found useful public relations advice, but no visible attribution to African PR knowledge producers. What does this reveal about whose expertise gets recognised?
By Hezron Ochiel
African universities, professional bodies, researchers and communication practitioners have produced extensive knowledge on public relations.
Their experience spans crisis communication, government relations, public health campaigns, stakeholder engagement and reputation management.
But when someone asks an AI assistant for advice on these subjects, whose expertise appears in the answer?
I decided to find out.
I tested ten common public relations questions across ChatGPT, Gemini and Claude. The questions covered PR strategy, stakeholder identification, media relations, communication effectiveness, crisis planning and misinformation management.
All three platforms produced useful guidance.
However, none of the responses visibly credited Hezron Insights, the Public Relations Society of Kenya (PRSK), the African Public Relations Association (APRA) or another clearly identifiable African PR source.
That was the most striking finding.
It does not mean the platforms never use African knowledge. It means the answers I examined offered no visible recognition of an African PR knowledge producer.
The distinction is important because a professional can benefit from an answer without knowing whose research, experience or ideas contributed to it.
How I conducted the review
I selected ten questions commonly asked by communication professionals, students and organisational leaders.
They covered the foundations of PR practice and practical questions about planning, implementation and evaluation.
Among them were:
- How do you develop a public relations strategy?
- How should an organisation identify and prioritise stakeholders?
- How can communication teams determine whether key messages reached their audiences?
- What should a crisis communication plan contain?
- How can PR demonstrate its contribution to organisational objectives?
The original prompts did not mention African organisations or ask specifically for African examples.
I wanted to examine whether African sources would appear naturally when someone sought general professional advice.
For each response, I considered its usefulness, the professional frameworks identified, the organisations mentioned and the sources visibly credited.
The three tools were reviewed as user-facing systems.
ChatGPT had web search enabled and displayed links. Gemini and Claude provided detailed answers but did not consistently identify the publications informing their guidance.
These were different retrieval and citation conditions, which limits direct comparisons between platforms.
Where no citations appeared, I recorded an absence of visible attribution. I did not assume the platform had relied on no external material.
What the three AI tools revealed
The responses showed a strong familiarity with established public relations practice.
ChatGPT’s visible sources allowed me to examine some of the references supporting its answers.
Gemini and Claude frequently organised their advice around familiar professional concepts and frameworks.
For example, Gemini described PR strategy as a process involving research, measurable objectives, audience identification, messages, tactics, resources and evaluation.
It also used the Mendelow Power–Interest Matrix to explain stakeholder mapping.
Claude provided comparable guidance while incorporating budget considerations, channel selection, contingency planning and performance measurement.
The answers were generally useful for someone seeking an introduction to these subjects.
The weakness lay in source recognition.
A professional framework might be identified by name, yet the user could not always see the publications or institutions from which the explanation had been drawn.
That difference matters when examining whether African knowledge receives recognition in AI-generated guidance.
International frameworks dominated the explanations
The tools repeatedly used established approaches associated with international professional practice.
These included the Public Relations Society of America’s definition of PR, RACE, SOSTAC, the Mendelow Matrix, the AMEC Integrated Evaluation Framework and the Barcelona Principles.
Their appearance was understandable.
These frameworks give practitioners recognisable ways to plan activities, identify stakeholders and assess communication performance.
They also offer convenient structures for explaining complex topics.
For example, a question about PR strategy can be answered using research, objectives, audiences, tactics and evaluation. A question about stakeholder mapping can be organised around influence and interest.
The frameworks make knowledge easier to explain and apply.
African communication professionals also possess extensive experience in these areas, including work carried out in complex institutional and community environments.
However, that experience is not always documented in publicly accessible guides, research papers or practical frameworks.
A successful communication intervention may become a conference presentation or workshop discussion without developing into a publication that others can study and reference.
The opportunity is to produce more accessible professional evidence grounded in African practice.
Useful answers, limited source transparency
The strongest responses concerned topics for which established guidance is widely available.
When asked about crisis planning, Gemini and Claude covered activation procedures, decision-making responsibilities, holding statements, designated spokespeople, monitoring and post-crisis review.
Their answers provided practical starting points.
The same pattern appeared in questions about communication measurement.
Both recognised that reach alone cannot establish whether a message has been understood or acted upon.
Gemini discussed exposure, recall and attitude change.
Claude distinguished message penetration, comprehension, recall, attitude change and behaviour.
These distinctions are consistent with established communication evaluation principles.
The AMEC framework, for example, encourages practitioners to examine outputs, audience out-takes, outcomes and organisational impact.
Yet knowing that an answer follows established professional thinking is different from knowing precisely which source produced the knowledge.
The limitation becomes more significant when an AI-generated response draws on ideas that may have originated in research, institutional experience or specialist publications.
A useful answer may reach thousands of people while the underlying knowledge producer receives little or no visible recognition.
Where African research could contribute
The questions about communication effectiveness were particularly revealing.
Consider a public institution that publishes information about a new service.
Its communication team can report the number of media stories, website visits and social media impressions.
Those figures provide evidence of exposure.
They do not establish whether the intended audience understood the announcement, believed the information or completed the required action.
Answering those questions may require surveys, interviews, administrative records or other evidence.
This is an area where African communication researchers can make substantial contributions.
Communication audits conducted in universities, hospitals, government institutions and development organisations can reveal how audiences actually receive and use information.
These studies may capture problems that general guidance cannot adequately explain, including language preferences, institutional hierarchies, digital access and differences between official communication channels and the channels audiences rely upon.
Publishing such findings would give other practitioners evidence they could examine and apply.
It would also create original material that journalists, researchers and professional knowledge platforms could reference.
African context appeared without African attribution
One response from Claude used a health training institution to illustrate stakeholder mapping and crisis communication.
The example referred to students, alumni, lecturers, regulators, health-sector employers and communities surrounding campuses.
It was relevant to the kind of institutional environment familiar in Kenya and elsewhere in Africa.
However, the response did not visibly connect that example to an African university, researcher, professional association or published source.
This illustrates an important distinction.
An AI answer can use an example that resembles an African setting without visibly recognising African expertise.
The appearance of local context is therefore insufficient evidence that local knowledge producers have been credited.
For researchers and communication professionals, that distinction deserves closer attention.
African sources became easier to find when named
I carried out additional searches for relevant African organisations and publications.
The results showed that professional material was available from Hezron Insights, the Public Relations Society of Kenya and the African Public Relations Association.
The material covered areas including professional development, reputation, communication measurement, government communication and misinformation.
The knowledge existed and could be located when the organisations were identified in the search.
This suggested two different forms of discoverability.
Prompted visibility occurs when a user already knows an organisation or expert and searches for that source by name.
Unprompted visibility occurs when the source appears in response to a general question, without the user specifying it.
The distinction matters.
A journalist seeking advice on crisis communication may have no reason to search for a particular African professional association if they have never heard of it.
Likewise, a student asking an AI assistant about PR measurement may encounter only the sources that the system chooses to display.
In my earlier article, How AI Search Is Changing Trust, Visibility, and Public Relations in Africa, I examined the wider implications of this development for African institutions and their digital reputation.
The present review adds a specific observation: relevant African sources were discoverable by name, but did not receive visible credit in the general PR answers tested.
That does not establish a systematic pattern across every platform or topic. It identifies a question deserving more rigorous research.
Why might African PR sources be missing?
The test did not reveal the technical reasons for the absence of visible African citations.
Several possible explanations warrant examination.
The first concerns how professional knowledge is published.
A professional body may organise excellent training on crisis communication but publish only a brief announcement about the event.
The announcement demonstrates that the training took place. It may provide little guidance to someone searching for an answer about crisis planning.
A detailed guide containing the training’s principles, examples and supporting evidence would serve a different purpose.
The second concerns accessibility.
Useful information may remain inside member-only platforms, consultancy reports, academic documents or poorly organised institutional websites.
Some restrictions are justified by confidentiality, copyright or commercial considerations.
Nevertheless, knowledge that cannot be accessed publicly is less available to readers and search systems.
The third concerns the strength of the wider publishing environment.
Established international organisations often maintain extensive libraries of definitions, research, templates and explanatory material.
Their publications may also receive numerous external references.
An African institution publishing one isolated article may be less discoverable than an organisation with many connected resources on the same subject.
These are plausible explanations, not findings established by the exploratory test.
Further investigation should examine which factors actually influence source selection and citation.
What African knowledge producers can do
The findings point towards a practical publishing challenge.
Professional associations, universities and communication experts should consider how their knowledge can serve people beyond those attending their events or using their services.
A workshop on misinformation could produce a practical response guide.
A communication audit could result in a publicly accessible research summary.
A professional conference could generate a collection of case studies examining what worked, what failed and why.
The resulting publications should identify the author, institution, date and evidence supporting the conclusions.
Where research is involved, readers should have enough methodological information to assess its reliability.
Related material should also be easy to locate.
An article on communication measurement, for example, could connect readers to research findings, practical reporting templates and relevant professional standards.
I examined these wider responsibilities in The Ultimate Guide to AI Search, Brand Reputation, and Public Relations in Africa.
The aim is to strengthen the usefulness and credibility of publicly available knowledge.
Whether these efforts improve AI citations should be tested rather than assumed.
How future reviews can improve the evidence
The ten-question exercise was exploratory. A larger study would need more consistent testing conditions.
The same prompts should be submitted across comparable platform settings, with web retrieval enabled where available.
Each test should record the exact question, platform, model, date, location, citation settings and sources displayed.
Researchers should also distinguish among direct citations, unlinked mentions, attribution through intermediary publishers and complete absence from the visible answer.
Prompted and unprompted searches should be assessed separately.
Repeating the questions over time would help establish whether the observed patterns remain stable.
A broader sample could also examine specific areas of practice, including public health communication, crisis management, government communication and stakeholder engagement.
Such work would provide stronger evidence about the circumstances under which African sources appear.
What the findings can and cannot establish
The review involved ten PR questions and three AI tools at one point in time.
It was not a representative study of all public relations knowledge available online.
The tools were also reviewed under different retrieval and citation conditions.
ChatGPT displayed web links, whereas Gemini and Claude did not consistently show the supporting sources.
The results therefore concern visible source recognition, rather than a complete account of the information each system used internally.
The test cannot establish deliberate exclusion of African sources.
Nor can it demonstrate that improvements to websites or publishing formats would necessarily cause a particular AI tool to cite an African institution.
The defensible finding is narrower: the responses reviewed provided useful professional advice without visibly crediting an identifiable African PR source, even though relevant African material could be found through direct searches.
That finding deserves further investigation.
From knowledge to recognition
The experience has reinforced a concern I have been exploring in my work on digital visibility.
African communication professionals possess expertise that can contribute to international understanding of public relations.
Their experience includes working with diverse communities, responding to crises, supporting public health programmes and managing communication in complex institutional environments.
Making that knowledge visible requires more deliberate documentation and publication.
Professional bodies can develop practical resources from their training. Universities can make research more accessible. Practitioners can publish evidence-based accounts of their work.
AI platforms also have a responsibility to improve the transparency and reliability of source recognition.
For now, the ten-question review leaves a question worth pursuing:
When people ask AI systems for professional guidance, how can they discover and recognise the African knowledge that helps answer those questions?
Hezron Ochiel is a strategic communications and public relations professional with more than 15 years of experience in journalism, digital communication and reputation management. He works at Kenya Medical Training College (KMTC) and is the founder of Hezron Insights, where he publishes research and analysis on strategic communication, AI visibility, SEO, GEO and digital authority.