Artificial Intelligence (A.I) In Public Relations: A Guide For Best Practice
A storm is brewing. While AI in PR may not have lived up to the hype since OpenAI launched ChatGPT in November 2022, change has come to our profession, and this is just the beginning.
As a result, the overview I give here is based on my experience and the latest findings from public relations trade organisations at the time of writing.
Please note – this is an abridged version of the full guide to AI in public relations available in the upcoming second edition of Making Waves:
Contents
Introduction
A storm is brewing. While AI in PR may not have lived up to the hype since OpenAI launched ChatGPT in November 2022, change has come to our profession, and this is just the beginning.
As a result, the overview I give here is based on my experience and the latest findings from public relations trade organisations at the time of writing. We are all very aware that AI is improving rapidly and to stay up to date it is essential to follow the latest guidance, blogs and newsletters from PR bodies like the CIPR and the PRCA.
My strong sense however is that the primary message in the first edition of Making Waves has even greater resonance post ChatGPT launch. That is, the role of a PR person at board level has never been more needed in the AI world. PRs are the eyes and ears and the gatekeeper helping protect companies from the ethical dangers of AI as well as harnessing its considerable powers to refine content and strategies.
The primary job of a PR person should not be to write press releases and manage media relations. That is a tactical job. The primary role is to build trust and relationships with the stakeholders that count to encourage them to deepen their relationship with your organisation. Media relations is one important element of this and the advent of AI does not mean the art of media relations is dying. Neither is the discipline of writing dying, AI cannot write copy as well as many of the top class writers in PR. Rather, AI is best used to refine content rather than create it. Nevertheless we must up our game beyond simply writing and focus on reputation and relationship building. Today, writers are still very much in demand as clients still require bespoke content. If clients rely on AI responses for, say thought leadership articles, they will have the same content or similar to everyone else. Moreover, writers are needed to deploy emotional intelligence to writing and editing in partnership with AI understanding, for example, finding exactly the tone the CEO wants to communicate.
But the challenge now for PRs is to more urgently elevate reputation management to the c-suite. AI can help you in this role like the internet and iPhone have helped transform our roles in the last 20 years. However, PRs not already at board level must skill up to be taken seriously by the c-suite. This is not just training in AI but in how they talk, dress and in their overall knowledge of business. This is far beyond simply media relations. For too long, PRs have lacked formal qualifications in PR, management skills and knowledge of business. Clearly a first point of call is to enroll with a PR organisation like the Chartered Institute of Public Relations or Public Relations Society of America and commit to a programme of learning, showing evidence you are upskilling.
But this overall lack of expertise and qualifications have conspired to become a barrier to entry to the c-suite for PRs historically. This has been to detriment of many maritime firms, who often lack PR expertise or a feel for marketing. Companies’ reputations are more vulnerable without a person on their senior team who is horizon scanning risks, prioritising stakeholder communication, building relationships with the media and listening to societal trends. As a result companies are not only at risk of reputational damage but they are missing growth opportunities.
Today there is little doubt that AI presents a profound risk to organisations and society. And this is where humans and PRs are needed more than ever. AI is actually not intelligent because it is not sentient, aware of context and does not think. And AI results do not have an ethical compass or understand a company’s values without comprehensive instruction. AI content depends on the quality of data the large language model (LLM) has been fed. Think of AI like an automated translation software, it anticipates words based on analysis of large data sets. The AI effectively parrots sequences of words without understanding their meaning. This is where the human, and the PR, must intervene, edit and cross check sources.
There is an inherent danger in being dazzled by the shiny new AI toy, and PRs must be disciplined to see it as a tool to assist them not surrender to. Anne Gregory the Professor Emeritus of Corporate Communication at the University of Huddersfield cautions: “Sentient intelligence such as empathy, humour, understanding how organisational leaders think, and ethical judgement cannot and should not be given over to AI.”
Taking a more sober view of AI, PRs should consider more closely the reputational risk of AI – deep fakes, copyright infringement, inaccurate responses, data sharing, privacy, misinformation, bias. There is also the issue of the vast energy use and how this could clash with a company’s commitment to sustainability. And specifically in maritime, AI creates cybersecurity threats, seafarer safety issues and challenges over regulation compliance. These are all issues a PR should be horizon scanning for and advising a maritime company about at c-suite level. Clearly any crisis planning now needs to be rigorously updated to take AI threats into consideration, examining how they can be prevented in the first place. And PRs should be advising on how a company should respond rapidly if a damaging deep fake of your CEO is circulating online.
Bias is further becoming a major concern for companies using AI. In the scenario of handling the analysis of junior CVs for example, the AI may have inherent bias towards people who have been employed before based on their class, race, gender, social background and the types of words they use. In this scenario without human intervention the company’s values could be compromised and someone who shows potential, from say a poorer background, could be ignored by the AI. PRs need to be alert to highlighting reputational risks like this to senior management.
The bottom line is that AI is exerting even more pressure on PR pros to graduate from more junior tasks to become senior strategic advisers. Overseeing AI is one vehicle PRs can use to achieve this. For this reason I believe time will show that AI was a force for good, driving a long overdue recalibration of PR rather than being an existential threat.
What is AI
AI as a phrase is misleading. It is artificial and at the moment not very intelligent. As discussed, AI is based on training an algorithm to manipulate a set of data called a large language model (LLM)
The vendors, like OpenAI, have claimed that the AI will eliminate much professional work and that the output is capable of demonstrating emotional and intellectual reasoning.
AI works by predicting words or phrases based on the data set or large language model it has been trained on. Although there has been massive hype about AI, there is limited evidence of anyone achieving anything like the 40% or more productivity gains promised by AI vendors like OpenAI, Google, Microsoft and Anthropic.
However, a picture is beginning to emerge of the tasks which have been replaced or assisted by AI tools.
Firstly ‘assisted by AI’ doing often repetitive or time consuming tasks such as making meeting notes and summaries with an AI like Otter.AI. And secondly ‘augmented by AI’ doing tasks that we would find difficult to do, such as content and image creation, fact checking, brainstorming and pitch ideas. AI’s like Canva, ChatGPT, Gemini and Claude allow us to be more effective because they enable us to do things we would find impossible or resource intensive.
So presently there is nothing disruptive or revolutionary in these tasks. AI is like an intern helping us do mundane tasks more efficiently. But like an intern is does require guidance when it gets things wrong and detail for how you want it to present information. Unlike some interns it only needs to be instructed once. AI is also a democratising force for those lower skilled and those in poor parts of the world as anyone with a web browser on the internet connection can access a LLM.
How to use AI
There are three ways to use a LLM such as a ChatGPT. You can access it for free, pay for a professional version or license it within an enterprise organisation.
There is a further option enabling organisations to build their own GPTs. It should be noted that Open AI launched a GPT store in January 2024 enabling third parties to publish their own custom GPT models. Its goal is to build an ecosystem around the ChatGPT platform similar to Apple’s app platform. More than two million apps have been published in the GPT store covering a range of applications such as editing image generation and research. The cost of app entry is low just $20 a month for a commercial ChatGPT account.
Slow adoption in PR
So far in public relations, according to PR thought leader Stephen Waddington, it is proving challenging to integrate AI into public relations work at scale. A survey published by Cision and PR week in January 2024 found that just 32% of PR’s were using AI frequently. This mirrors a report by the Institute of Public Relations in February 2024 which found high levels of anxiety around hallucination – where the AI invents an inaccurate answer – leading to misinformation in practice.
However, a turning point may be in sight. The Boston Consultant Group found consultants using ChatGPT completed tasks 25% quicker and produced 40% higher results than those without the tool. Goldman Sachs takes more conservative position and believes AI will have a 7% impact on global productivity over the long-term. Waddington says that AI performance enhancement is relative. He says: “It has greater impact on improving performance and effectiveness of low performers than high performers and it’s not yet clear what tasks AI is best suited to support. As a result the effect is the ‘jagged frontier’ and can only be determined by using AI in PR practice.”
There are literally thousands of AIs that can be used by PRs. The main tasks we are seeing AI used for in PR, according to a Global CommsTech report in March 2023, are media and social media monitoring, content creation and databases.
The largest opportunity was seen in analytics while the biggest threat is cyber security and privacy. The most important emerging skills are measurement and evaluation, followed by data science, behavioural insights and video and image creation.
AI Regulation
Maritime firms and PRs must keep a close eye on AI regulation. In maritime this is made far more complex by the global nature of operations and the fact different countries and even US states have their own regulations.
In the United States, regulatory and compliance frameworks have lagged behind AI’s rise. However this is set to change as AI is catching the attention of federal and state regulators. Notably in the absence of comprehensive federal legislation on AI, there is now a growing patchwork of various current and proposed AI regulatory frameworks at the state and local level. State lawmakers are considering a diverse array of AI legislation, with hundreds of bills introduced in 2025. These proposals mainly fall into the following categories:
- comprehensive consumer protection legislation
- sector-specific legislation on automated decision-making
- chatbot regulation, requiring chatbot providers to inform users they are not interacting with a human and impose liability on chatbot providers for misleading chatbot communications.
- generative AI transparency requirements,
- AI data center and energy usage requirements,
- Frontier model public safety legislation.
In the UK, the government has adopted a principles based non-statutory framework for regulating AI focused on leveraging existing laws and regulations rather than introducing new AI specific legislation.
The European Union however is taking a more interventionist approach and has taken steps to regulate AI systems with the Artificial Intelligence Act, called the most comprehensive regulatory framework for AI systems in the world. The AI Act introduces a risk based classification system for AI applications categorising them as an unacceptable, high, limited, minimal which determine their regulatory treatment.
AI risks
1 Unacceptable Risk
This means that an AI system poses a threat to people and their fundamental rights, contravening EU values. This includes biometric identification such as facial recognition in public places, social scoring such as ranking people on personal characteristics, socio-economic factors or behaviour. The regulation will ban these types of AI.
2 High Risk
Significant threats to health and safety or fundamental rights of a person. Systems that may have a negative impact on fundamental rights will be classed as high risk and divided into two categories: those used in products covered by the EU’s product safety legislation (e.g. aviation or medical devices), and those that will need to be registered in a database (e.g. systems used for law enforcement or migration, asylum, and border control management).
3 Limited Risk/Transparency Risk
Users must be made aware of when they’re interacting with AI, so applications susceptible to manipulation, like chatbots or deepfakes, must be transparent and comply with EU copyright law. This covers apps that make it possible to manipulate images, sound or video. These will be regulated by security, transparency and quality standards.
4 Minimal Risk
Systems used for video games or spam filers, Most applications are expected to fall in this category and will not be regulated,. Standard voluntary codes of conduct are recommended.
Who is covered by the AI Act
The regulation applies to all 27 member states of the European Union. The regulation will further apply to companies located outside the EU if the AI system is offered in the EU market or impacts people living in the EU.
When does it come into force?
The EU AI Act is something of a moveable feast. Here’s a timeline of the most notable dates:
- April 2021 – The European Commission proposes the first EU regulatory framework for AI.
- March 2024 –Parliament adopts the Artificial Intelligence Act.
- May 2024 – The Council approves the AI Act.
- June 2024 – European Union lawmakers sign the act.
- Aug. 1, 2024 – The EU AI Act enters into force across all 27 EU Member States.
- Aug. 1, 2026 – After a two-year implementation period, the regulation will be fully applicable.
AI search rocks publishing world
AI is rocking publishers in a way not seen since the dawn of internet caught them off guard. They are now catching up. And for PRs it is now clear that AI search is set to disrupt company reputation in the way search engines have for the last 20 years. It has often been said in marketing ‘You are who Google says you are’. Well we can now extend that to ChatGPT, Gemini, Perplexity and other LLMs. Research by PR agency Hard Numbers published in December 2024 looked at how LLMs influence perceptions of quality, trust, innovation, and value in brand communications. The research revealed that traditional media significantly influences AI-generated content about brand reputation, with editorial media driving 61% of responses.
Perhaps surprisingly LLMs show a high-level trust in owned media, often accepting corporate claims without scepticism. It provides more evidence why corporate communications urgently need to start treating LLMs as stakeholders. As a result it will now be very important to manage high quality informed AI search responses for clients and ensure they are kept fresh.
Two primary points from a publisher perspective.
First how publishers are to be compensated for AI using content for training. LLMs must be updated to ensure they are contemporary and publishers are data rich so Vendors want access to their huge archives. For example OpenAI brokered a deal with Associated Press in 2023 to license its content as training data. This was followed by a deal with the Financial Times in April 2024. However, other publishers are much more protective with more than 500 news publishers including the New York Times, Reuters and the Washington Post installing a blocker that prevents their content from being collected and used to train ChatGPT.
Second publishers and AI vendors are recognising an opportunity to monetise search. A Journalism and Technology Trends and Prediction 2025 report from the Reuters Institute Study found publishers will be prioritising building relationships with AI platforms such as ChatGPT and Perplexity to drive referral traffic. Both AI companies have in turn been courting high-quality content in return for publisher citations and/or money.
While in the early stages at the time of writing this relationship between publishers, AI vendors and search is likely to be a major factor in the media and AI worlds in the years to come. Not only will it be a very big revenue raiser for AI and publishers it will also be a key source of traffic for media websites. Studies so far have shown that AI search draws from news websites and company websites while social media does not factor at all.
PRs will need to stay on top of how this relationship between publishers and AI search develops to manage reputation and correct any inaccuracies.
However, in the wider publishing world AI remains a highly contentious force with fierce concerns expressed over breach of copyright and plagiarism.
This was evidenced by the Authors Guild, a trade association for US authors, which has sued OpenAI in New York for unlawfully training is LLM‘s using copyrighted content generated by its members. The case claims that the unauthorised use of work is identity theft on a grand scale.