By Leanne Link , Dave Erickson
In the past few months, we have seen a great deal of interest in generative artificial intelligence (GAI). People are experimenting with GAI applications like ChatGPT, and businesses are thinking about its impact on customer experience, accounting, marketing, and more. Given how quickly technology evolves, it can be difficult right now to judge what is speculative and what is actually implementable and valuable.
We are now at a point where government leaders should seriously consider how to prepare internal data to derive maximum value from GAI, and how to use GAI to facilitate better citizen and employee experiences.
GAI itself is only as good as the data it was trained on
In its current state, GAI can produce impressive content, dialogue, images, and more. But those results are only as relevant as the data the tool was trained on. When the training datasets—providing representations of knowledge in large language models (LLMs)—are based on publicly available data on the Internet, they generate a limited range of answers. GAIs based on public data are often prone to hallucinations—presenting incorrect information as accurate.
On the other hand, when GAI is used with an agency's internal data, it can dramatically speed up mission outcomes, improve citizen services, and better connect government knowledge workers such as analysts and cybersecurity professionals with the right people at the right time. data are linked. Why? Because the agency data adds the necessary context .
The combination of GAI and private agency data has a force multiplier effect. The naive solution is to bake private data into the models themselves; however, the complexity and cost of training or fine-tuning AI models—multiplied by the number of government domains and interaction points—becomes untenable. Instead, the same questions asked of LLM can first be used in Elastic's AI-powered search capabilities, where the most relevant fact-based answers based on your internal data can be found.
This domain-specific context that your data brings to GAI can make the output more accurate, relevant, and better suited to your task. The premise of "bring your own data" is that your data is stored on a unified data platform and can be accessed and found in one place.
What about privacy and security?
Especially for the public sector, you don't want to mix highly sensitive data with publicly accessible GAI or any system where you don't control your own data. Any search query sent to a publicly available GAI product such as ChatGPT is consumed by the model, meaning your internal data is no longer internal. Even if your organization isn't officially using GAI as part of your tech stack, it's a safe bet that your employees will be using it anyway.
Help ensure your internal data is in the right hands by strategically integrating GAI with your proprietary data in a way that your IT team can control and gain insight. Otherwise, you risk allowing employees to inadvertently put your sensitive data into public GAI services like ChatGPT without you being able to ensure its security. Ideally, you can integrate your proprietary data into a platform designed to handle sensitive information, where you retain full control over your own data and enable role-based access control (RBAC). Read more below.
Accelerate task impact using GAI
Data is one of the most strategic assets owned by public sector organizations today. When your data is unified and stored on one platform - where it can leverage GAI and search techniques - the real world impact can be profound, offering benefits including:
Personalized access to public services
Imagine a citizen seeking to apply for public housing services. The application process involves several steps and forms that vary by need and location. Simply listing general information on a web page would be complex and may not address a citizen's unique situation. On the other hand, when institutions bring their own data to GAI, citizens can find information and instructions tailored to their individual circumstances. This highly relevant information has the potential to reduce the complexities that often prevent people from accessing basic services.
Simplify the citizen experience
Or, to take another example: You've been called to serve as a juror and need to know what's going to happen next. where do you need to go How long will it take? Have you been selected as a juror? Do your judges allow cell phones in court? Using your data, GAI can simplify and personalize this complex information, potentially improving the citizen experience and building trust with government services and leaders.
Accurate Investigation and Intelligence
For law enforcement and the intelligence community, democratizing access to the right data in real time is critical. This is especially true when multiple organizations are collaborating on a project -- using different databases of information in different formats. Being able to find answers across data types and sources with a single GAI query has the potential to improve the speed and accuracy of results, reduce manual and time-consuming work, and ensure that everyone who needs it can work with the same accurate dataset.
Improve employee productivity
When you integrate GAI with domain-specific context, you can help your internal teams quickly find the information they need to help them do their jobs. Fast queries across multiple datasets and formats can deliver highly relevant information in real time—avoiding the need to laboriously (and mind-numbingly) comb through documents or siled databases. In most cases, the information your team is looking for won't be found on the public internet or in AI model training sets, so it's important to provide a GAI-powered tool to quickly find proprietary information so your employees don't turn to Public tools that can compromise your data security.
When employees spend less time on fruitless searches and manual data correlation, you remove more sources of friction from their day, paving the way for better job satisfaction and engagement, especially when you When resources are tight at the outset.
GAI + Elasticsearch + your internal data
The Elasticsearch platform can be a powerful tool when you consider how to integrate your institution's data with GAI. It allows you to ingest all types of data, store it economically, access it anywhere, and integrate it with GAI transformer models.
Elastic has been democratizing search for more than a decade, and we've been investing in artificial intelligence and machine learning (ML) for much of that time. That's why we just launched the Elasticsearch Relevance Engine (ESRE) to help our customers find relevant answers to their questions through AI and ML on the Elasticsearch platform.
What is the Elasticsearch Relation Engine (ESRE)?
ESRE combines the best of AI with Elastic's text search, providing the ability to integrate with large language models (LLM). It's accessible through a simple, unified API that the Elastic community already trusts, so developers can start using it immediately to improve search relevance.
In other words, you can now connect your own GAI models or third-party GAI models directly to the data you store in the Elasticsearch platform. This allows you to leverage the power of GAI and domain-specific data to generate answers that are accurate, relevant, actionable, and secure.
To learn more about ESRE, read the release blog .
Why choose Elasticsearch for GAI and private data?
1) Unified data storage and democratized access . You can affordably store all your data in the Elasticsearch platform to democratize access, findability, and insight. Once your data is in the platform, you can use it for other use cases such as threat hunting and infrastructure monitoring.
2) Ability to find mission-critical answers :
- Accurate : You'll base the answers you'll get from the GAI and your own data on facts relevant to the task -- not hallucinations.
- Relevance : By using proprietary data in Elasticsearch, you can avoid repeatedly retraining the LLM on internal data, saving you time and training costs, and ensuring your information is always up to date.
- Actionable : The Elasticsearch platform democratizes access to data and insights, enabling your teams to collaborate anywhere and make decisions in real time.
- Security : Not every employee should be able to access every document, and for data sovereignty purposes, some data needs to reside in a specific location. Elasticsearch allows you to restrict data access to certain roles within your organization while still retaining the ability to search across your entire data store.
3) Cost-effective implementation . Thanks to decades of optimizations in information retrieval, Elasticsearch interactively presents knowledge to GAI in a way that is orders of magnitude more CPU-efficient than extracting the same knowledge from large trained or fine-tuned language models. Some estimates suggest that semantic retrieval is five times more efficient than just using ChatGPT 3.5 or 250 times the CPU cost of GPT-4.
How much value GAI can create for your organization depends on your data and how unified and accessible it is. If your data is spread across multiple tools and teams, you may lack the context and content needed to make GAI highly relevant to your mission goals. The Elasticsearch Platform serves as a single data store for all your organization's data, and a centralized starting point for collaboration, AI insights, and automation.
Next step
- Learn more about Elasticsearch and GAI .
- Learn how to implement Elasticsearch and AI for privacy-first use cases from a technical perspective .
- Contact Elasticsearch public sector experts to discuss how AI can bring value to your agency's mission.