Beyond Buzzwords: making ethical AI a reality for health brands

Introduction

Artificial Intelligence (AI) has been around for a while now, and whilst recent advances in generative models have brought it into the spotlight, they rest on decades of research, development, and real-world application. Today, AI is transforming how we work, connect, and communicate across nearly every industry, including healthcare. Tools that once seemed futuristic are now within reach for healthcare brands of all sizes. Yet for many, the idea of adopting AI can still feel overwhelming. 

In fact, some of the most effective applications are simple, easy to implement, low risk, and highly impactful. The most important step in any AI journey is the first one. And with the right approach and guidance, it can be far more achievable than you think. 

Ethics, regulation, and accountability in AI 

While AI may be a new tool in the healthcare communications toolbox, the existing frameworks that guide compliance, data privacy, and promotional practices still apply. The current standards that ensure accuracy, balance, and patient confidentiality remain especially relevant in the context of AI. By using these processes and ethical checks to AI-generated work, brands can confidently explore innovation without compromising trust. 

At the same time, the pace of AI development is moving faster than regulatory safeguards. There is a noticeable gap in AI-specific regulation around the practical risks and ethical considerations. However, it’s important to note that while this isn’t necessarily a barrier to what can be done, it is creating a confidence gap. It’s this lack of confidence that’s leaving many healthcare brands hesitant to move forward. This uncertainty, however, presents a unique opportunity for forward-thinking brands that are willing to take the first steps. By proactively engaging with trade bodies and shaping internal best practices, healthcare brands can help set the standard before formal policies are in place. This can make them stand out as leaders in a market still finding its footing. We understand the creative ambition of brand teams and the regulatory rigour of approval teams - we’ve worked closely with both. This means we can help you bring bold ideas to life in compliant, credible ways. Whether it’s shaping a patient-centred experience, building a future-proof brand platform, or optimising workflows with AI, we’ll help you deliver meaningful impact without friction.

AI doesn’t have to feel risky 

Most people’s first experience with AI has likely come from personal use through tools like Siri, Google Gemini, ChatGPT, and other consumer-facing platforms. These may be easy and convenient to use, but questionable when it comes to privacy. These services typically collect large volumes of user data, retain conversations, and use that data to retrain their models. Understandably, this fuels the perception that AI can’t be trusted with anything sensitive. 

However, this perception doesn’t reflect the reality of how AI can be used in professional contexts. 

There is a wide spectrum of privacy and security options available, including: 

• Switching off logging 

• Disabling data retention 

• Preventing inputs from being used for training 

• Designing closed-loop, custom-trained systems to maintain data confidentiality 

It’s even possible to create a fully private enclave of AI capability, built around your brand’s data and regulatory obligations, and accessible only to your teams and trusted partners. 

Having delivered a range of real-world AI applications across healthcare and pharma, we’re well positioned to help you navigate the options. Together, we can find a solution that aligns with your challenge, your budget, and the level of control you need. We don’t believe in going big just because you can. We believe in making smart, intentional choices that deliver the right balance of precision, privacy, and practicality. 

Inclusive by design

It’s no secret that AI models carry bias. If you were to ask a generative AI model to create an image of a doctor, you’d likely get a white, middle-aged man in a lab coat. After all, the data these models are trained on often over-represent certain demographics and stereotypes. While we can’t fundamentally change the models themselves, we can influence the outcomes they produce. 

AI models are designed to ‘regress to the mean’, meaning they will give you the most statistically average answer possible. And because they’re trained to be agreeable, they’ll often default to the safest, most common representation they can find. While that might make outputs look convincing, they are not necessarily accurate or fair. That’s why we believe inclusion needs to be built in from the start. 

That’s where prompt engineering comes in. Prompt engineering is the practice of crafting specific, intentional instructions (or prompts) to guide an AI model towards producing more relevant, ethical, or inclusive outputs. For example, instead of simply asking for an image of ‘a doctor’, we can help shift the outputs toward more accurate representations by including specific prompts, such as ‘a female neurosurgeon in her early 40s’. 

Of course, prompt engineering is just one approach. Ensuring inclusive outputs also means addressing AI bias at the source. This can be achieved by using representative, diverse training data, avoiding overfitting, and critically evaluating how generalisations are made. Different AI applications require different safeguards, and the right combination depends on the use in each case. 

In many ways, this isn’t so different from what we already do in healthcare communications. Every day, we make careful decisions about how we represent patients, HCPs, and products in a way that’s thoughtful, ethical, and inclusive. Using AI just means applying the same level of care, consideration, and critical thinking in a new context. 

With the right prompts, checks, and principles, we’ll work with you to ensure AI outputs reflect the diversity and inclusivity your brand stands for, and that the people you’re speaking to feel accurately and respectfully represented. 

The human-AI partnership 

Balance is key, particularly in healthcare communications where clarity, accuracy, and empathy are critical. That’s why AI works best when paired with human insight. AI can generate, retrieve, or reframe information, but it is human oversight that provides context and final judgement. Indeed, that’s the principle behind a human-in-the-loop model: AI supports the process, but humans are ultimately in control. This leads to greater efficiency without sacrificing integrity or nuance. 

With the right approach, AI can make data dramatically more accessible. Intuitive searching and document retrieval can turn static archives into dynamic, discoverable resources. In fact, as AI becomes more capable, healthcare organisations will be able to unlock the full value of their internal data by curating high-quality, well-organised repositories that AI can tap into effectively. This shift could accelerate product development, improve customer experience, and guide smarter decision-making across the organisation. 

Large Language Models (LLMs) excel at adapting content for different formats, audiences, and reading levels. But while they’re great at rephrasing, they’re less skilled at interpreting nuance, particularly cultural and emotional nuance. What sounds empathetic in one language or region may feel cold or confusing in another. This is where human expertise matters - ensuring that tone, intent, and context all land in the right way. 

When approached with intention, AI can act as both a creative and strategic partner by enhancing efficiency and enabling more personalised, inclusive communication. 

That includes being transparent when AI is used, so people understand how content is created and can trust the process behind it. Used thoughtfully, AI can help teams communicate with greater empathy, consistency, and care. 

Future-proofing AI 

The AI landscape is evolving rapidly, which is why future-proofing isn’t just about adopting the latest tools, but about staying adaptable in the face of rapid change. What works today might be obsolete tomorrow, and the biggest shifts may be the ones we don’t see coming. 

We help clients build long-term capability by focusing on generalised, transferable skills and technologies. By embedding flexible systems, encouraging continuous learning, and anticipating change, we help teams stay resilient whatever the future brings. For organisations with the available resources, we recommend bringing in expert partners who can navigate complexity quickly and effectively, and help accelerate progress. 

We understand that AI isn’t a magic solution that will reinvent your business or create an entirely new source of value. Instead, AI can scale and amplify your existing strengths and specialisms, whether in product development, content creation, or customer engagement. Future-proofing isn’t about certainty. It’s about readiness, agility, and having the right mindset (and support) to move forward with confidence. 

Sustainable AI 

The environmental impact of AI is a significant, yet often overlooked, concern. Training and running large-scale AI operations can be incredibly energy-intensive, requiring vast computational resources and generating significant carbon emissions. 

In most cases, the more resource-intensive the model, the more costly it is too. Thus, smarter, more efficient choices are better for both the planet and the bottom line. In practical terms, this means: 

• Leveraging smaller, more efficient language models where possible 

• Using prompt engineering in place of energy-intensive fine-tuning 

• Adapting existing technologies rather than developing entirely new ones from scratch

By making intentional choices, we reduce the environmental impact of our AI usage without compromising the results our clients expect. 

Be brave. Take the first step. 

It’s natural to feel hesitant in a landscape filled with technical jargon, regulatory grey areas, and evolving ethical concerns. But don’t let complexity stand in the way of innovation. Be brave. Taking a simple, focused approach is often the smartest way to begin. 

We’re here to help you use AI in a way that works for you - efficiently, ethically, and effectively. Our experience across creative and regulatory spaces means we know how to make things happen, without compromising on compliance or ambition. 

If you have a brand challenge, a creative block, or an opportunity you’re unsure how to tackle, don’t hesitate to contact us. We’ll work with you to explore how AI, applied with care and expertise, can help you achieve your communication goals. Take that first step and reach out to Claire: claire.dobbs@solrishealth.com for a chat about your AI challenges. 

Disclaimer 
We use AI to push creative boundaries, deliver impactful work and improve efficiency. We are committed to transparency around AI use and will always use it in line with the terms of client contracts. We use pre-vetted AI tools, audited under the Mission AI Acceptable Use and Guidance Policy, to meet our intellectual property, data protection and client confidentiality standards. If any work is in part generated by AI, we will let you know and advise you of any limitations in fidelity, resolution, adaptability, reproducibility or licensing.