In my 30 years of experience in this industry, I’ve never seen anything quite like the hype, misunderstanding, and raw power of a technology like we’re seeing with AI.
Let’s address some of these issues today.
We’ve been talking to a lot of people about AI and how this groundbreaking technology can accelerate data extraction and automation.
We’ve been privy to these discussions because BIS created Grooper to utilize technology for the document space more than a decade ago. This industry has always had a foundation in AI. In the mid-1990s, I worked with mathematical algorithms that are the basis for the AI tech we’re all enamored with today (Large Language Models).
Speaking of LLMs, these are the products that everyone is worried about when it comes to privacy. There’s been a lot of miscommunication and misunderstanding around how LLMs “use” your data.
How is Grooper Securely Using AI?
First, there is a difference between using these models (LLMs) when you connect via an API (Application Programming Interface) versus using a web-based application like ChatGPT. This is especially true if you are using a FREE version of the ChatGPT or any other chat client.
But Grooper integrates with AI directly through APIs. This means we have enterprise-level security built into our integrations. You get an API key and a secure endpoint. This is the same level of security and encryption that your online banking uses when you log into your bank’s website.
Grooper’s AI Does Not Keep Your Data
However, encryption isn’t the only issue people are concerned about. The more significant issue is the misbelief that LLMs take and use your data to train and get “better.”
This is not the case with daily use of an LLM through an API. The process of getting information back from an LLM is called inference. This is different and distinct from the initial training that LLMs use to be built.
If you use a free chatbot, like ChatGPT, the company (OpenAI) might use your data to train its models in the future. This is because if you use a free product, the company gets something back for your product usage—free data from you.
However, when you use an API (like Grooper does), OpenAI and others do not keep your data. In some instances, your data might be kept for 30 days for any violations of usage terms, but you can opt out of these systems with all of the leading vendors. Some LLM providers make it even more clear that they never keep or use your data.
Better to Be Safe than Sorry
We highly recommend that you push your legal teams to examine the privacy policies of any LLM you are considering using. I’ve encountered some large companies that have paid to create a “safe” version of their LLM integration, which was largely unneeded. But sometimes, fear leads companies to make decisions.
All of this reminds me when companies started to become very interested in the internet. A lot of companies were scared and made comments like, “I’ll never put my data online.” Yet here we are.
LLMs are just the latest technology to come down the pike that is “scary.” That’s because even the people who created them can’t really tell you how they work. Not really.
That doesn’t mean they aren’t safe, however. If you’ve been following these recent newsletters, you’ve seen the power that can be harnessed and the money and time that can be saved by embracing LLMs.
We’ve spent considerable time designing our AI integrations with safety, accuracy, and governance in mind.
Grooper Gives You Many LLM Options
However, if you really can’t use a cloud-based LLM for one reason or another, you still have another option with Grooper. Grooper supports any LLM with an API connection via OpenAI’s standard.
And all of the open-source LLMs can be run with an OpenAI API on top of them. Meta, Google, Anthropic, and hundreds of others have state-of-the-art open-source LLMs that can be run in your data center without any reliance on the cloud, and Grooper can fully utilize them.
And BIS can even sell you the hardware and software for the LLM to run on. We’re one of Dell’s biggest partners in the state of Oklahoma.
The short version is that LLMs are safe when used through Grooper. Your data is not being used without your knowledge. And ignoring the value of AI is going to hurt you in the end.
If you’d like to learn more, please reach out to me!
I’m happy to show you how far we’ve come, and how you can use AI today with your company’s resources!
