writing a chatgpt policy for your organization

By Dr. Jim Castagnera, Esq.

Partner, Portum Group International, LLC

Artificial Intelligence made a quantum leap forward in November 2022 when Open AI released ChatGPT.  While most experts place us five to ten years away from “The Singularity” --- that moment when an AI is intellectual equal to a human being and also self-aware --- ChatGPT in the words of Elon Musk is “scary good.” There’s no avoiding the impact of ChatGPT.  Your employees, your customers, your clients, your students… they’re all enamored of it, never mind that it may eventually replace their jobs and end their careers.  Consequently, every organization needs a policy on ChatGPT right now.

An AI Bill of Rights

“The Blueprint for an AI Bill of Rights is a set of five principles and associated practices to help guide the design, use, and deployment of automated systems to protect the rights of the American public in the age of artificial intelligence. Developed through extensive consultation with the American public, these principles are a blueprint for building and deploying automated systems that are aligned with democratic values and protect civil rights, civil liberties, and privacy.”

https://www.whitehouse.gov/ostp/ai-bill-of-rights/what-is-the-blueprint-for-an-ai-bill-of-rights/ 

The components of an AI Bill of Rights, as proposed by the White House include:

• Safe and Effective Systems: 

“Automated systems should be developed with consultation from diverse communities, stakeholders, and domain experts to identify concerns, risks, and potential impacts of the system. 

“Systems should undergo pre-deployment testing, risk identification and mitigation, and ongoing monitoring that demonstrate they are safe and effective based on their intended use, mitigation of unsafe outcomes including those beyond the intended use, and adherence to domain-specific standards”

• Algorithmic-Discrimination Protections

“Designers, developers, and deployers of automated systems should take proactive and continuous measures to protect individuals and communities from algorithmic discrimination and to use and design systems in an equitable way. 

“This protection should include proactive equity assessments as part of the system design, use of representative data and protection against proxies for demographic features, ensuring accessibility for people with disabilities in design and development, pre-deployment and ongoing disparity testing and mitigation, and clear organizational oversight.”

• Data Privacy

“Designers, developers, and deployers of automated systems should seek your permission and respect your decisions regarding collection, use, access, transfer, and deletion of your data in appropriate ways and to the greatest extent possible; where not possible, alternative privacy by design safeguards should be used. 

“Systems should not employ user experience and design decisions that obfuscate user choice or burden users with defaults that are privacy invasive.”

• Notice and Explanation

“Designers, developers, and deployers of automated systems should provide generally accessible plain language documentation including clear descriptions of the overall system functioning and the role automation plays, notice that such systems are in use, the individual or organization responsible for the system, and explanations of outcomes that are clear, timely, and accessible.”

• Human alternatives, Considerations and Fallback

“You should be able to opt out, where appropriate, and have access to a person who can quickly consider and remedy problems you encounter.”

With regard to this last point, there are two obvious possibilities:

1. Human in the Loop

“A human is assisted by a machine. In this model, the human is doing the decision making and the machine is providing only decision support or partial automation of some decisions, or parts of decisions. This is often referred to as intelligence amplification (IA).”

Source: Harvard Business Review, November 10, 2021, https://hbr.org/2021/11/managing-ai-decision-making-tools 

2. Human on the Loop

“Here, the machine is assisted by a human. The machine makes the micro-decisions, but the human reviews the decision outcomes and can adjust rules and parameters for future decisions. In a more advanced set-up, the machine also recommends parameters or rule changes that are then approved by a human.”

- Ibid.

Why You Need a ChatGPT Policy Now!

 Among the ways organizations are using ChatGPT include:

• Writing templates for online content. 

• Customer service correspondence. 

• Writing code. 

• Writing sales pitches. 

• Summarizing long reports. 

• Analyze business trends. 

https://www.thecorporategovernanceinstitute.com/insights/news-analysis/policy-for-chatgpt-may-be-crucial-as-ai-surges-in-popularity/

• Language Translation

• Chatbot

• Research

https://trainual.com/template/chatgpt-policy

• Fact Checking

• First Drafts

• Editing

• Brainstorming

https://www.legal.io/articles/5429675/You-ll-Probably-Need-a-ChatGPT-Company-Policy

Despite being “scary good,” ChatGPT has serious shortcomings:  

• “That bank does not keep up with any news cycle. The most recent information could be months, if not years, old. This means any ChatCPT-produced content could ignore the most recent relevant events. 

• “The information bank can include biased sources. ChatCPT could misinterpret these as hard facts and present them as such. 

• “The bank may contain sensitive data, which ChatGPT could deem fair game for widespread publishing. If organizations use ChatGPT for published content, they become liable.”

• https://www.thecorporategovernanceinstitute.com/insights/news-analysis/policy-for-chatgpt-may-be-crucial-as-ai-surges-in-popularity/

Issues to consider covering in your policy include:

1. Data security and privacy

2. Discrimination, bias, and disparate treatment

3. Transparency and comprehensibility

4. Responsibility and accountability

5. Other ethical considerations

6. Monitoring and correcting

7. Employee training and supervision

https://www.thecorporategovernanceinstitute.com/insights/news-analysis/policy-for-chatgpt-may-be-crucial-as-ai-surges-in-popularity/

A Model ChatGPT Policy

This policy comes to us compliments of an organization that generously has developed it and made it available for general adaptation.  Kairoi requires the following: “Simply copy and paste the below content into your preferred text processor. Our only ask is that you acknowledge and respect this template's license by adding the following line in a clearly visible part of your document: ‘Based on Template ChatGPT Use Policy by Kairoi Ltd (2023) / kairoi.uk / CC-BY 4.0’.”

https://github.com/KairoiAI/Resources/blob/main/Template-ChatGPT-policy.md

So here goes:

Introduction

• Artificial Intelligence (AI) tools are becoming increasingly useful and accessible. In particular, generative AI tools have become a part of many professionals’ workflows. Generative AI tools include those that produce text or images in response to queries. They are an accessible way to interact with large language models (LLMs), AI systems trained on enormous amounts of data. In that training process, LLMs require huge compute resources. In practice, text-to-image tools can be used by publishers to create visuals, and artists to aid them through the creative process. Similarly, chatbots based on LLMs can help write blogposts, emails and reports. ChatGPT is one such LLM-based chatbot.

• ChatGPT serves as an interface to interact with an LLM that is trained on enormous amounts of data. When inputting a prompt in ChatGPT, we are basically interacting with all the data it was trained on. Through different machine learning techniques, ChatGPT is also trained to predict the next word in a sentence, which allows its responses to sound natural and even compelling. However, ChatGPT can provide factually incorrect information, is easily manipulated, and is further trained by the prompts users provide.

• At [org name], we do not encourage the usage of ChatGPT, but we acknowledge that staff may value integrating the tool into their workflows. [Org name] will not provide you with login credentials but, if you choose to use ChatGPT, you may do so by signing up with a personal email and mobile number at https://chat.openai.com/, notwithstanding the present policy. It is expressly forbidden for you to create an account to use ChatGPT with [org name] email addresses or phone numbers.

• The present policy sets out general guidelines, as well as forbidden and acceptable uses of ChatGPT, to ensure the safety of staff, the company, and our partners, clients and suppliers. Whilst [org name] cannot monitor your use of ChatGPT on personal devices, we do have a [Computer Use Policy] whereby your use of devices of [org name]’s property are subject to relevant monitoring procedures. The present ChatGPT Use Policy complements and in no way supersedes [org name]’s [Computer Use Policy], and applies to all staff employed by or seconded to [org name].

General Guidelines

• ChatGPT is a tool, similar to using Google to browse the internet, or Word to write documents

• When you type into ChatGPT, you are giving information to the system’s developers at OpenAI

• Always ask yourself whether ChatGPT is the best tool for the task at hand (ask your colleagues or supervisor if you're not sure)

• Each time you start a new chat on ChatGPT, you are starting “from scratch” in the sense that it cannot reference other chats you had with it

• ChatGPT can provide inaccurate information, and all its responses must be carefully checked for errors

• ChatGPT is best used to feedback on one’s own work, and not to scrutinize others’ efforts

• When seeking feedback on text or code, it’s safest to provide only snippets to ensure you are minimizing the amount of information divulged

• If you don’t want something to be in the public domain, then don’t type it into ChatGPT!

Acceptable Uses

• Ask for a spell-check to ensure proper grammar (for example: "Please check the grammar of this sentence: "these processes must be separated"")

• Ask for better ways to word a sentence (for example: “What's a better way of saying "our service is faster and more efficient"?”)

• Ask for the outline of reports if it is the first report of the sort (for example: “What's a useful outline of a recruitment report?”)

• Ask for the outline of a blogpost for inspiration (for example: I need to write a blogpost introducing privacy-enhancing technologies – how could it be structured?”)

• Ask for feedback on snippets of code or help with Excel formulas (for example: "I am using Excel. Column A has the price "$10" and column B has the number of avocados offered at this rate ("2"). How do I get the price per avocado in column C?")

Forbidden Uses

• You may not sign up to ChatGPT using company credentials (work email and phone number)

• You may not divulge sensitive information, such as passwords, addresses, bank details or phone numbers

• You may not type out or paste draft contracts, such as employment contracts or contracts with clients or providers

• You may not type out or paste documents that may contain business-sensitive data, such as draft annual reports, business cases and accounting reports

• You may not seek feedback on job applicants’ CVs by pasting them in in part or in full

• You may not type out or paste proprietary code, documents or other information that is not intended for the public eye

• You may not type out or paste content that is subject to copyright licenses that do not allow for their indiscriminate sharing

• Failure to comply with this section of the present policy will result in disciplinary action and can lead to termination of employment.

Join Me For a Deeper Dive into Getting Your Arms Around ChatGPT

How to Write a ChatGPT Policy for Your Employees

Date : June 20, 2023| EST : 01:00 PM| PST : 10:00 AM| Duration : 60 Minutes

https://assentglobal.us/webinar/3106/How-to-Write-a-ChatGPT-Policy-for-Your-Employees