If you’ve ever typed a prompt into ChatGPT and thought, “That’s not quite what I meant,” or hesitated before pasting anything sensitive into an AI tool, you’re asking the right questions.
As a freelance medical writer, I’ve wondered the same things:
- Can AI really support med comms work?
- Are free tools safe to use on client projects?
- How do I avoid hallucinated citations and still deliver high-quality work?
My take? I believe AI can help, but only if we use it with care, context, and a healthy dose of trial-and-error. Personally, I treat AI as a behind-the-scenes assistant: helpful for brainstorming, structuring drafts, or unblocking tricky sentences. But it’s not a replacement for deep thinking or editorial judgment. And if a client prefers I don’t use AI at all, I respect that completely. This post is a transparent look at how I use AI tools thoughtfully, and always with the goal of delivering clearer, stronger work.
Inside, I’ll share:
- The free and affordable AI tools I actually use, and where they’ve proven valuable in my day-to-day writing.
- Why I stopped using some tools, and the ones I’m still keeping an eye on.
- My freelance-friendly checklist for vetting any AI platform to save you time, help you stay compliant, and protect the trust you’ve built with clients.
Let’s dig into what works (and what doesn’t) when it comes to free and low-cost AI tools for medical writing.
The AI Tools I Use (and How They Fit into My Workflow)
ChatGPT / Claude
These two have become near-daily companions in my workflow. I don’t use them to write for me, but to help me think more clearly and work more efficiently.
I often use them for:
- Ideation: getting unstuck at the start of a project by asking for angles, themes, or alternative ways to frame the same message.
- Restructuring: improving the logical flow of a slide deck or rewriting a paragraph to make it more digestible.
- Simplifying trial data: I’ll paste in the results section of a press release and ask for a plain English explanation, then cross-check the output carefully.
- Titles and section headers: when I need something sharper or more client-friendly, I’ll ask for 3–5 alternatives to work from.
I use both ChatGPT and Claude regularly, but I’ve noticed some key differences. Claude is better at helping me with writing tasks like suggesting a logical flow for blog posts or comparing drafts. It handles large amounts of text more smoothly and lets me upload documents directly. Claude then generates what it calls “artefacts”, i.e. summarised or analysed outputs that sit alongside the source data, which I find really helpful when I need to work with or reflect on longform content.
That said, these tools aren’t magic. Hallucinations are still a real issue, particularly if you’re asking for detailed or up-to-date scientific information. They still lack nuance and can swing too formal or too vague.
Top Tip: When prompting, frame your request as if speaking to a junior colleague. For example, “Explain these results to a healthcare journalist,” or “Suggest five ways to present these outcomes in a slide.”
To level up your AI prompting skills as a medical writer, check out my guide: Mastering the Art of AI Prompts for Medical Writers. It includes a proven prompt formula and real medical writing examples to save you hours and sharpen your output.
For a broader take on how this tool is reshaping our work (and why it still needs a human touch), read: ChatGPT for Medical Writing: My Thoughts on How It’s Changing the Game.
Otter.ai
Otter has been a staple in my toolkit for years, especially when recording client calls or expert interviews (always with permission). It automatically transcribes meetings and lets me focus on the conversation rather than scribbling notes.
The transcripts are rarely perfect. Technical terms often come out garbled, and speaker attribution can get muddled. But for a general summary or to pull out key quotes, it’s more than sufficient.
Compared to Zoom or Google Meet’s transcription features, I’ve found Otter more user-friendly and easier to organise by project. That said, some clients now prefer to use their own platforms, and that’s something to accommodate.
Pricing (as of 11 November 2025):
- Free Plan: Includes 300 transcription minutes/month and 3 lifetime file uploads.
- Pro Plan: $8.33/user/month with 1,200 minutes and 10 file uploads/month.
The free or Pro plan are usually enough for freelancers, unless you’re transcribing long interviews regularly. The free plan is a good starting point if you’re curious.
Perplexity
Perplexity is my go-to when I want to get a fast overview of a topic. For example, treatments in a particular therapeutic area or recent trial activity. It’s not as structured as Scite, which I trialled for a while, but it’s free and can be surprisingly effective.
That said, it doesn’t handle complex queries particularly well, and like many AI tools, it sometimes pulls outdated information or mentions sources which are not credible. That’s one reason I stopped paying for Scite: while the interface was excellent, the cost didn’t justify the marginal gains for my kind of work.
Here’s a recent example of what came up when I searched “folic acid in pregnancy” on Perplexity:
With any AI search tool, the golden rule is: always verify the sources yourself. While it’s a great springboard, I always double-check claims before citing them. Even if it links to a trusted source like a PubMed entry or a guideline, I go there directly and never take citations at face value.
AI Tools I’m Curious About (But Haven’t Fully Adopted Yet)
Like many freelancers, I’m always keeping an eye out for tools that could streamline my workflow. A few that have caught my attention recently include:
Notion: The All-in-One Organiser
Notion is endlessly customisable: dashboard, project tracker, calendar… you name it. I’ve seen some medical writers use it to manage client content calendars, store templates, publish websites or track progress across multiple deliverables. The appeal is the “all-in-one” aspect, but I haven’t committed yet because of the learning curve.
Paperpal: Academic-Style Writing Support
Paperpal is marketed as an AI writing assistant for academic and scientific content, making it a potential fit for medical writers needing help with citation generation, paraphrasing, translation, and manuscript readiness. I haven’t used it in live projects yet, but the ability to work directly in Word or Google Docs is tempting.
Pricing (as of November 2025):
- Free plan:
- 200 language editing suggestions/month
- 5 uses/day of generative writing tools (rewrite, predictive writing, translation, AI review)
- Limited chat with PDFs and citation features
- 7,000 words/month for plagiarism checks
- Prime plan: €11.40/month (billed annually at €137)
- Unlimited editing suggestions and writing tools
- Advanced PDF chat with citation exports
- 10,000 words/month plagiarism checker
- Unlimited journal readiness checks
A free version is a good starting point for trying out Paperpal’s key features.
Jenni AI: Built-In Citations
Jenni combines AI-powered writing support with citation integration. What makes it stand out is its ability to draw on your uploaded PDFs, which could be helpful when working with dense source material. It’s very geared toward academics, but I can see potential for content-heavy med comms projects too.
Pricing (as of November 2025):
- Free Plan:
- 200 words/day of AI autocomplete and edits
- 10 PDF uploads
- 10 AI chat messages/day
- Partial document export
- Unlimited citations in 2,600+ styles
- Unlimited Plan: $12/month
- All features unlimited (autocomplete, edits, PDF uploads, chat, exports)
- Full document export
- Priority support and access to new features
Beloga: Collaborative Knowledge Workspace
Beloga positions itself as a collaborative knowledge assistant that pulls from sources like Google Drive, Google Scholar, and Notion all at once. It’s designed to help teams write from shared, searchable content, which could be especially useful for working with clients who want to centralise background material, source links, brand guidance, and key messaging in one place.
While I haven’t used it in a live client project yet, I can see its potential for streamlining feedback loops or co-creating content when multiple stakeholders are involved. If you’re working with internal teams or agency partners who often say, “I’ll just drop it in the shared folder,” Beloga might be a more structured alternative.
For now, it’s probably more relevant to collaborative or larger-scale environments, but it’s definitely on my radar as something to test with clients who value shared access to research.
Pricing (as of November 2025):
- Free Plan:
- 1 workspace
- Rate-limited searches
- GPT-4o-mini model access
- Community support
- Pro Plan: $20/month
- Unlimited workspaces and searches
- Access to advanced models (GPT-4o, Sonnet-3.5, Gemini-1.5, etc.)
- Specialized data sources and 1:1 productivity coaching
- Priority support and early access to new features
Vespper: AI for Regulatory Drafting and Review
Vespper is built specifically for life-science regulatory work. It’s designed to automate the drafting of complex documents like CSRs, SOPs, and INDs, and support regulatory research (it can search across internal documents and databases like EMA, FDA, MAUDE, and PubMed). It offers features like contradiction detection, regulatory-aware editing, and traceable AI outputs. It’s definitely more enterprise-focused (suited for agencies or pharma clients) and priced accordingly, but I’m keeping an eye on how this might evolve for smaller teams or freelancers.
Pricing: Not publicly listed (demo required)
Why I Haven’t Jumped In Yet
While these tools are promising, I’m cautious for a few reasons:
- Data privacy: I work with confidential materials, so I’m selective about where I paste content, especially if tools don’t have clear compliance policies.
- Workflow fit: My current tools are simple and work well. Adding something new needs to solve a clear problem, not create extra admin.
- Return on investment: Many of these platforms come with monthly fees, and without a clear time-saving or quality improvement, it’s hard to justify the spend.
That said, I’d love to hear from others who are using these tools day-to-day.
How I Vet AI Tools: A Freelance Writer’s Checklist
With so many AI tools popping up, it’s tempting to try everything. But I’ve learned (sometimes the hard way) that not every shiny new platform is worth the time, cost, or risk.
Here’s the checklist I use before adding any tool to my workflow:
Cost vs. Value
Freelancing means watching both time and money. I ask myself:
Does this tool save me time or improve quality enough to justify the cost?
If I’m using it daily or it replaces two other apps, it’s usually worth it. But if it’s just “nice to have,” I hold off.
Accuracy and Hallucination Risk
No matter how polished it sounds, AI still gets things wrong. Some tools invent sources or misinterpret data, which is a no-go in medical writing. I test with real examples and check how often I have to fact-check or reword.
If I’m spending more time cleaning up than I would writing from scratch, it’s not worth using.
Data Privacy and Client Trust
This is non-negotiable. I never input client names, unpublished data, or sensitive materials into tools that don’t clearly outline how they handle and store user content.
If a client asks, I want to be able to say with confidence how (and where) their information is being used.
Integration Into Existing Workflows
The best tools fit into how I already work (not the other way around). I ask myself:
- Does it work with Google Docs or Word?
- Can I copy/paste easily without losing formatting?
- Does it speed up what I’m already doing or create extra steps?
If a tool adds friction, even if it’s powerful, it rarely sticks.
Transparency and Explainability
Some tools give answers or summaries without showing where the information came from. That’s a red flag.
I look for tools that:
- Cite sources (ideally with live links)
- Let me inspect the “why” behind the output
- Make it easy to trace and verify information
This is especially important when I’m creating content that needs clinical, regulatory, or academic backing.
This checklist helps me avoid jumping on every trend and instead focus on tools that actually support quality writing for me and my clients.
Final Thoughts: What AI Can (and Can’t) Do for Medical Writers
If you’re a freelancer exploring free or budget-friendly AI tools, you don’t need an expensive subscription to benefit. Many of the tools I use daily (like ChatGPT, Otter, and Perplexity) have free or low-cost tiers that offer real value when used intentionally.
That said, here’s what I’ve learned:
- If a client prefers a fully human-authored approach, that’s what I’ll always deliver.
- Be curious but cautious. AI isn’t a shortcut but a support system. I don’t want to be relying on AI too heavily, so I’ve stuck to my rule: use AI to shape, not substitute your thinking.
- Test tools on real work. Play around with live projects, like drafting section titles, summarising trial outcomes, or organising meeting notes, to see what genuinely saves you time.
- Trust and accuracy come first. No AI tool replaces editorial judgement. I never paste confidential info, I always fact-check outputs, and I stay transparent with clients about how I work.
Free and low-cost AI tools could exactly be what you need, as long as you know what to use them for, and what to leave to your own expertise.
To learn how to prompt like a pro, get my guide: Mastering the Art of AI Prompts for Medical Writers.
It includes a proven structure and real examples for sharper, faster AI results.
You might also like:
- Effective AI Prompts for Writers: A Practical Guide
- ChatGPT for Medical Writing: My Thoughts on How It’s Changing the Game
FAQs: Free and affordable AI tools for medical writing
Are free AI tools safe to use for client work as a medical writer?
It depends on how you use them. Most free tools are not HIPAA or GDPR compliant, so avoid inputting client names, unpublished data, or anything under NDA. If in doubt, anonymise your prompts or work offline with general examples. Also, be transparent with clients if you’re using AI in any part of your process.
Can AI tools help with regulatory or journal writing?
In some cases, yes, but tread carefully. Tools like Vespper and Paperpal are built with traceability and compliance in mind. They can help with structure, editing, and even cross-referencing. That said, AI should never replace your expertise in interpreting guidance documents, summarising data, or aligning with client expectations. Always double-check the output, especially if it’s going into anything high-stakes.
Do journals or clients consider AI-assisted writing plagiarism?
Most journals and clients are still catching up with policy, but the key difference is whether the AI is generating original content or copying existing material. Using AI for idea generation, drafting outlines, or simplifying text is fine — as long as you’re not copying and pasting AI-generated content without reviewing or editing it.
For more on ethical AI use in med comms, see my blog: ChatGPT for Medical Writing: My Thoughts on How It’s Changing the Game.
