How to use ChatGPT in FP&A with Data Privacy?
Christian M.
Finance Transformation Senior Manager at Kraft Heinz | AI in Finance Professor | Conference Speaker | LinkedIn Learning Instructor
Hi!
Today, I want to bring you 2 ways to use ChatGPT with enhanced data security and data privacy.
Why?
But one of their main concerns is data privacy.
How to address it?
I have 2 ways:
1) ChatGPT + Python + G. Colab
2) ChatGPT Enterprise
I'll give you the details of it.
1) ChatGPT + Python + G. Colab
Earlier this week I posted a way to create dashboards with ChatGPT.
And some FP&A professionals had concerns like "what are the commercial sensitivity implications of uploading data", or "the data becoming part of OpenAI's training set and effectively become publicly available".
So how to address this?
The process is simple:
Step 1: Ask ChatGPT to generate Python code to do the type of analysis you need.
You can use my framework below to make it easier!
Step 2: Copy-Paste this code into Google Colab
Google Colab is like Google Docs for programming!
领英推荐
It's very easy to use and learn and you don't need to download anything!
Also, it is free with some premium tiers available.
You can access it here.
Step 3: Add your real data into Colab
I have explained it with more detail in this video.
Btw, I mainly use G. Colab when I teach to beginners because it is so easy to learn and use!
But if your company prefer Microsoft Products, then I'd recommend Visual Studio.
This is one of the topics we covered on our Advanced Chat GPT Course and I want to give you a $100 discount on it.
This link will expire by Tuesday 13th August.
2) ChatGPT Enterprise
Now the second way!
Built for businesses, ChatGPT Enterprise offers organizations the ability to use ChatGPT with controls, deployment tools, and speed required to make your entire organization more productive.
Your workspace admins control how long your data is retained.
Any deleted conversations are removed from our systems within 30 days, unless OpenAI is legally required to retain them.
There is more detail here in OpenAI's website.
And I created an article with more details on how to use this method.
AI and treasury transformation: treasurer turned coach, helping multinational treasury teams to improve cash flow by millions and reduce workload by 20%+ | Experienced FTSE100 Treasurer | Speaker
7 个月Well done on addressing this as for most large organisations the idea of sharing any sensitive information, let alone non public financial information with a public LLM is an absolute no go. For that reason many organisations are using local hosted versions of llama or slms specifically for the purpose. At this stage I’d be tempted to advise to use Microsoft copilot over open AI given open AI’s patchy governance record and the regular resignations there, although some functionality is lost by using Microsoft
Projektmanager bei BMW Financial Services / Central & Southeastern Europe
7 个月Christian M. Hi Christian, love the technique. However, with method 1 the data privacy discussion in my view moves from ChatGPT to Google Colab (for uploading files), data still remaining in the US. European FP&A teams may continue to be reluctant to the 2 proposed solutions if data is stored/processed in US, which poses the data privacy concerns. How well do you methods apply in the EU? Any additional aspects to consider? Love the exciting new horizons with AI, best Andreas