Applying FDA regulations intelligently is key to your medical device business success
DINA LAKAYAN
01
What is it
ChatGPT for regulatory
Large language model AIs are perfect for (in vitro diagnostic) regulations
Everyone knows ChatGPT. If you’ve tried it for your RA questions though, you know that its data is outdated since 2023 and its answers are often incorrect.
Built on OpenAI technology
The engine of ChatGPT with the right knowledge
ChatFDA augments everyone’s favourite chatbot with up-to-date FDA regulations, the latest guidance documents and applicable standards.
Your data is safe
Input your company’s IP with confidence
One drawback of the consumer ChatGPT is that your data may be used for further training. With ChatFDA, your data is not shared with anyone and not stored by OpenAI.
News
ChatFDA released
ChatMDR and ChatIVDR 3.1 with Reasoning model
Benchmark of the Fast, Smart and Reasoning ChatMDR models
ChatMDR API available in beta An API, Application Programming Interface, is a way to connect…
ChatMDR 3 and introducing ChatIVDR
ChatMDR maintenance
ChatMDR maintenance Sunday, May 26, 20:00-22:00 CET On Sunday night, we’re going to have to…
02
Value add
ChatFDA promise
From the experience of CREATE RA
Innovative strategies
A strong medical device regulatory strategy is key to your business for the USA. Get it right and you can get ahead of the competition and spend your resources on innovation. Get it wrong, though, and you can spend thousands of hours on unnecessary paperwork and testing.
That said, a lot of RA work is tactical rather than strategic. Looking up facts, writing rationales and updating documents. That is where ChatFDA can do a lot of the heavy lifting.
Use your scarce resources intelligently
Don’t let your experts do routine
Since the introduction of MDR and IVDR in Europe, good RA people and consultants are hard to find. And their hourly rates are like those of lawyers. Are you going to let those people spend hours searching the FDA website and guidance documents?
Of course not! What you want is them to formulate the plan and then do the reference searching, fact checking and text writing to a computer.
From Art to Science
How ChatFDA works
A large-language model (LLM) like ChatGPT is trained on millions of documents. It is a neural network modelled after the human brain that, from all these documents, knows that there is a connection between cows and grass and milk, for example. But it is not specific to a niche of human knowledge such as the FDA medical device regulation – just like most people, all of whom understand cow-grass-milk, don’t either.
ChatFDA is an augmented version of ChatGPT that has access to the state of the art FDA CFR21 part 800, FDA guidances and standard documents. It’s the AI equivalent of a regulatory affairs expert – one with very good memory, willing to work on the weekend and not accepting job offers from other companies 🙂
ChatFDA is a chatbot like you know from ChatGPT. It is also based on the same technology by the inventors OpenAI. The big differences between ChatGPT and ChatFDA are, that ChatFDA uses additional, detailed and specifically prepared data around the contents of the 21CFR part 800 law, plus FDA guidance documents and ISO standards that ChatGPT does not have. Also, the creativity of the chatbot has been dialed down, to give you accurate results without hallucination. This is the process where a generative AI chatbot starts to ‘make stuff up’. That’s the last thing you want for your regulatory rationales!
Finally, at ChatFDA, we truly believe in keeping your data safe. ChatGPT got some bad press coverage for using the data you put in to train future model versions. ChatFDA does not do that. In other words, your sensitive company and product data that you may put in, are safe and never used for training the next AI model. Why is this relevant? You don’t want your competitor to go in, ask a question about one of their products that is similar to yours and they see your data!
03
About Us
Who we are
ChatFDA is a collaboration between CREATE and Morse Digital Media
Combining decades of experience in regulatory affairs and digital media and software
The brains behind ChatFDA are those of Dina Lakayan, founder and CEO of CREATE. “People ask me: will AI take over the regulatory profession? To which I say: probably, one day. We don’t have all the advanced reasoning skills in the computer yet. It can now be a bit brute-force and not ‘actual intelligence’, if you will. But ChatFDA can save you the time of a junior RA consultant or, even better, let your senior people use their time more effectively.”
The ChatFDA technology was designed by Maarten ter Mors of Morse Digital Media. “When I first heard about generative AI and ChatGPT, my immediate thought was that this is going to do so much good for regulatory! Generative AIs are good with text, searching, and summarizing. Exactly what you need when you have a bookshelf full of regulations and guidance documents. However, the standard ChatGPT is too generic and steals your data. That’s why I teamed up with Dina to train our own AI. One that knows the FDA 21CFR 800 and keeps your secrets safe.”
