The Do’s and Dont’s of Using AI (Like ChatGPT) For Your Business

Everywhere you look, people are talking about AI – specifically ChatGPT – and how businesses can use it. Your business has followed suit and made an account, but where do you start? It is true this technology has exciting capabilities. However, it is important to know how to use it without breaching legal and ethical guidelines. In this article, we will set out how you can use AI like ChatGPT sensibly and effectively to help promote efficiency in your business without breaching the law. 

What is ChatGPT?

ChatGPT is one of the most common forms of AI that businesses are increasingly using. It operates like an advanced chatbot, designed to understand and generate human-like text. ChatGPT generates its answers based on vast datasets that it has learned from “training”. It does not use the internet, so the GPT-3.5 model (the free version) only has access to information up to September 2021.

How Are Businesses Using ChatGPT?

Businesses continue discovering new ways of using ChatGPT, especially as technology advances. It has gained popularity by being able to generate content within seconds.

The Do’s

Fact Check

There are limitations to the data ChatGPT uses to generate its content. Currently, GPT-3.5 has only been trained on data available up to September 2021. This means that responses can often include incorrect or inaccurate data, facts, or figures. Subsequently, this results in misleading or unreliable information in a business document. Additionally, ChatGPT is not advanced enough to interpret complex data, such as legislation. For instance, ChatGPT is notorious for suggesting legislation or regulations that simply do not exist.

Therefore, you must fact-check any content or document you use ChatGPT to produce. 

Use It To Inspire, Not to Create

Letting ChatGPT take the wheel when drafting your documents can be tempting. While it is helpful to point you in the right direction or give you a headstart, you should not use ChatGPT should mindlessly. Automated drafting using ChatGPT lacks human touch and intuition. It may not capture the exact tone, sentiment, or empathy needed for certain business documents that require a personal touch. It is best to use its responses as inspiration for the work you can tailor according to your business style. 

Keep It Simple

There is no doubt ChatGPT has emerged as a powerful disruptor in the business world. However, it is designed to understand and generate responses based on the information it receives from a user. If this information is confusing or ambiguous, the response you receive will be limited. It is, therefore, important to keep your communication on ChatGPT clear, direct, and simple. This enhances its ability to provide the most relevant and appropriate response. 

De-Identify Information

As mentioned, ChatGPT uses previous user input to generate future content. Accordingly, when using ChatGPT for your business, you must not use any information that could identify your clients, employees, customers or competitors, as this could raise issues with privacy and confidentiality (discussed below). Instead, you should ensure that you de-identify your input to ChatGPT. De-identifying data refers to removing or modifying information that could identify a company or individual so that the data or text is anonymous. 

AI Art: Your Legal Considerations Factsheet

This fact sheet outlines your rights and obligations as an AI artist regarding intellectual property and copyright.

Download Now

The Don’ts

Input Personal or Confidential Information

One of the key risks businesses face when using ChatGPT is the risk of breaching privacy laws and confidentiality obligations. This can happen when a business includes sensitive information in their input to AI, which then uses this data as “training” for future responses. It is, therefore, best practice to never include sensitive or identifiable information when using ChatGPT. 

Privacy Laws

Under Australian privacy laws, personal information may only be used or disclosed because it was collected (the primary purpose). Exceptions, such as individual consent or legal requirements, are necessary for using or disclosing information for a secondary purpose. Typically, the primary purpose is to provide services to customers, and using technologies to support that purpose is allowed. However, you will still have obligations to ensure any third parties treat personal information securely. Using customer information to train third-party generative AIs or providing such information to generative AIs that do not adequately protect that information would likely violate Australian privacy law.

Confidentiality Obligations

Generally, contracts will contain an obligation of confidentiality. Disclosing ‘confidential information’ to a third party will likely constitute a breach of a party’s contractual confidentiality obligations unless that disclosure falls within one of the exceptions listed in the contract (if any). In the absence of such an exception, disclosure would constitute a breach of contract for which damages may be payable.

Claim Ownership

The law in Australia is still playing catch up with AI. While cases have begun trickling into the courts, the law remains generally unsettled as to who owns the work generated by AI. The likely position in Australia is that copyright protection is unlikely unless a human ‘author’ has applied some ‘creative spark’, ‘independent intellectual effort’ or ‘skill and judgement’ and not be copied from another work. This is where the above point comes into play: you should only use ChatGPT to inspire, not create. Without retaining ownership of the output generated by ChatGPT, you will have limited recourse regarding copyright protection.

Key Takeaways

Although ChatGPT is still in its early stages, it is already proven to be a powerful tool for businesses to produce work quicker and more efficiently. However, with new technology comes new risks. Businesses must exercise caution with what they input into ChatGPT and the responses they receive back. Fact-checking and de-identifying input are both key to ensuring the work produced is accurate, reliable and complies with privacy and confidentiality obligations. The work must also be edited by someone who can add their personal touch, creativity and intelligence to put themselves in the best position from an intellectual property perspective. 

If you are looking for advice on ChatGPT within a business context, our experienced artificial intelligence lawyers can assist as part of our LegalVision membership. For a low monthly fee, you will have unlimited access to lawyers to answer your questions and draft and review your documents. Call us today on 1300 544 755 or visit our membership page.

Frequently Asked Questions