In the blink of an eye, ChatGPT is already smarter than many of us. It’s incredible how quickly this innovative technology has taken the world by storm. Launched by OpenAI in November 2022, the online tool has quickly garnered attention for its immediate and detailed responses across a vast array of online knowledge. You can type in any question and within seconds, you can watch the tool type out a well-written answer to your question. This technology isn’t going away, and there are numerous implications to its use.

At its core, ChatGPT — Generative Pre-trained Transformer — is a chatbot built on top of a family of large language models fine-tuned with supervised and reinforcement learning techniques. If you ask ChatGPT what it is, the tool will respond and tell you that it is: “capable of generating human-like text based on the input it is given. The model is trained on a large corpus of text data and can generate responses to questions, summarize long texts, write stories and much more.”

ChatGPT resembles an online customer service chat icon that you click on and ask questions. Yet, it is much more. It immediately scans a massive array of online information to develop its answers. Then, it uses the questions and information you provide to fine-tune its future answers to other questions. This is called “reinforcement learning from human feedback.” Scary, right?

Now that you know what it is and are rightfully a bit wary and outright petrified, why should one care? Many people, including county staff, are using it to quickly solve everyday problems. Surveying local government IT professionals, I have heard a litany of great use cases including:

• Shorting research time
• Helping developers to write more efficient code and batch scripts
• Providing an automated way to write infrastructure routines
• Helping improve written communications (letters, memos, emails, etc.)
• Providing templates for policies, job descriptions, etc.
• Generating interview questions
• Responding diplomatically in a heated conversation.
• Creating employee and manager surveys
• Writing resolutions
• Translating language

While there are tremendous benefits, there are challenges everyone should be aware of. Challenges include:

• Data privacy – one needs to be very careful about what you are putting into ChatGPT. In a recent article in CSO, “Sensitive data currently makes up 11 percent of what employees paste into ChatGPT, with the average company leaking sensitive data to ChatGPT hundreds of times each week, according to Cyberhaven.” Confidential information or personally identifiable information should not be part of the text you include. Further, ChatGPT has already succumbed to sensitive data exposure. ChatGPT Exposed Payment Card Data of Subscribers (govinfosecurity.com)
• Data accuracy – Answers are only as good as the information available. Fact checking is important, because some of the information returned may be based on non-factual information.
• Data relevancy – ChatGPT only includes internet data from fall of 2021. I expect this to change (if it hasn’t already). But for now, it is important to be aware of this limitation.
• Misuse – potential misuse by malicious and criminal actors is already occurring. Hackers are already using the bot to create more convincing malware or phishing scams and the model could be tasked with creating and spreading misinformation. And then there is the educational concern where students are using ChatGPT to write their papers; thereby, eliminating the need for student research and critical thinking.
• Critical thinking – Just as important is the potential elimination of critical thinking. As we all know, critical thinking is vital in resolution and problem solving, as well as in perfecting vision and implementing creative ideas. Will ChatGPT eliminate that crucial step in problem solving? Will it create bias? We need to be cognizant of both of those real outcomes.

Counties’ use of ChatGPT is in its infancy. Even with the challenges, the benefits are significant. It is safe to say that certain strategies should be considered. Here are a few that should be reviewed and implemented with support from your county IT leadership:

• Develop a county policy on the use of AI (artificial intelligence) that is responsible and ethical. Content should include the county position on:
  • Privacy and data security
  • Transparency and accountability:
  • Fairness and bias:
  • Informed consent
  • Responsible use
  • More information clarifying each of these areas can be found here. 
• Educate staff on ChatGPT and how it can be used in a positive way, but also pitfalls
• For technical readers, implement ChatGPT for county solution development through the Azure OpenAI Service.

NACo will continue to watch as this innovative technology evolves. For now, there are three areas to keep an eye on:

• How other countries are responding to ChatGPT. As of this writing, Italy has banned all Italian residents from accessing ChatGPT, claiming that ChatGPT has violated European Union’s General Data Protection Regulation (GDPR). While GDPR applies to European citizens, there are various state data privacy laws that are like GDPR. My perspective is that there will be national discussions on ChatGPT.
• The frenetic pace of improvements of newer versions of ChatGPT. ChatGPT 4 is already out for subscribers to the paid version. It includes the ability to interpret images and respond in written form on how to use the image or what the image stands for.
• Active discussions on whether AI should be paused. Even the research experts and well-known leading figures in technology are saying that a pause makes sense in order to give some time for the development of appropriate guardrails.

Readers may want to consider the following question: Will ChatGPT take your job? As Satya Nadella executive chairman and CEO of Microsoft claims "this technology is going to reshape pretty much every software category." Considering whether ChatGPT will replace jobs is a valid question.

What do you think? Have you or will you use ChatGPT?