Our Chicago Data & Analytics team is building a community for interest for high-level leaders in the data space to discuss hot topics and collaborate around opportunities and challenges they share. Most recently, they discussed how AI is evolving the data space. Here are some key takeaways from our most recent meeting.
In our most recent meeting, the group enjoyed a beautiful July afternoon on the rooftop deck at DePaul University Chicago’s downtown campus for an event focused on the promise of AI tools (specifically large language models such as ChatGPT) and their practical implications in the data world.
We considered a few central ideas as a group, each one’s dialogue led by a different member.
Topic 1: What are you currently doing with these types of AI tools? Where do you see realistic potential for AI in your company or industry?
We found that data executives in this group are using chatbots with generative AI for customer onboarding, writing proposals and providing proof of concepts. There are two additional areas where we saw many of our members experimenting with AI:
- Data Processing: Using AI tools to automate mundane reporting tasks and enable operators to convert and process information rapidly.
- Sales Optimization: Using historical data to make product suggestions.
These local leaders felt more confident using new, more broadly accessible AI tools like ChatGPT, Dall-E, Copilot and Soundraw for personal needs than enterprise-level ones in their current state. One participant cited that using ChatGPT for a DIY bathroom remodel was a more efficient way to get instructions than Google. The group agreed there’s a need for training on how to interact with these tools to produce reliable results.
As these technologies advance to become more reliable and adoptable, executives saw potential in implementing them at the enterprise-level in the future in the following ways:
- Microsoft Copilot for tedious reporting tasks.
- Synthesia for real-time translation during presentations.
- ChatGPT for deciphering notes, identifying pain points and creating roadmaps.
Topic 2: What are the risks, lessons learned, or worries around ChatGPT?
Next, we moved on to one of today’s hot-button topics – ChatGPT. Many attendees found the popular chatbot to be an unreliable tool because there’s no way to trace its answers back to the specific data points that led to the output. And, it sometimes generates made-up, nonsensical responses (or “hallucinations”). This also makes using tools like ChatGPT with HIPAA-protected or otherwise private data risky.
On top of this, executives in the medical and finance fields – where complying with regulations and consistent documentation are critical – voiced concern about the fact that it doesn’t always yield the exact same answer even if receiving the same question again.
Overall, the group agreed that ChatGPT has the potential to realize new productivity levels and success in the data world, but its current capabilities are not dependable and require considerable human guidance to be effective.
Everyone agreed this topic is growing at a tremendous pace, and it is easy to get caught up in the headlines and hype. They also discovered that none of them are “behind” or “losing the AI race.” All are learning, experimenting and looking forward to envisioning what AI might mean to them and their companies. Discussions like this one help them both explore the possibilities of AI while also remaining grounded in the realities of today.
The group gathers once each quarter to break down topics such as data fabric, data privacy, data quality, cloud-native capabilities, real-world data science, artificial intelligence, machine learning and more. We share real use cases around these topics and how our members are creating business value through information.