Everyone’s talking about it. Chat GPT is the generative AI technology that interacts with users in a conversational way. And it has made quite the impact in the world of business. Buzzfeed has recently laid off 12% of its staff to incorporate using Chat GPT more frequently for content creation. Social media and mainstream media is awash with opinions about its efficacy and potential uses.
But new technology is often met with a dose of scepticism and Chat GPT is no exception. Before you incorporate Chat GPT into your business strategy, here are a few of the legal risks with the current format.
Data breaches
If you allow your employees to use Chat GPT in the course of business, you need to be wary of potential data breaches.
The technology is constantly learning and it learns from the input that users provide. If your staff input sensitive data, then Chat GPT may use that confidential information to answer another user’s question. Lawyers for Amazon have recently warned employees against this very practice.
You can’t control the ways in which Chat GPT uses the data that you input. But you can put in place a policy to control the ways in which your employees use Chat GPT. If you allow your staff to use the technology in the course of business, be very clear which information is confidential and sensitive and continually keep your staff up to date with best practice for using generative AI.
Intellectual Property issues
There’s an inherent IP risk involved in using Chat GPT. The technology collects data by scraping websites for information. It has also ‘learned’ material from a range of books, textbooks and other written materials.
However, it provides answers without ever citing its sources. A general dictum of intellectual property law is that ownership of the content vests in the person who created it. Replicating work in the way that Chat GPT does edges very close to the line of stealing.
At the moment, it’s free to use Chat GPT. But the future model will be “pay-by-generated-word”. That means that Chat GPT will effectively profit from other people’s work. If you’re using Chat GPT’s output, then you may also face IP challenges from the ‘true owner’ of the work.
Don’t forget that this problem cuts both ways. If you have content or graphics that you want to protect from the content-gobbling machine of Chat GPT, then you should register your work for copyright or trademark.
Plagiarism
Chat GPT is a tempting option for students, either at school, college or university. Rather than going to all the trouble of reading books, researching and thinking for themselves, students can simply type essay titles into a chatbot.
While it’s unlikely that students would submit an essay entirely generated by AI, it still takes away the key aim of learning; to train your brain to think original thoughts. But that is perhaps an ethical issue rather than a legal one.
Most universities now have in place technology to detect plagiarism (such as Turnitin) and content that has been produced through AI (such as GPTZeroX). There are continually evolving programmes to detect AI generated content, which learning environments may need to incorporate more heavily into their marking policies to reduce the prevalence of plagiarism now that Chat GPT is available.
Inaccuracies
While Chat GPT has been trained on as accurate information as possible, it either doesn’t always interpret the information correctly, or it relies on false or outdated information. In fact, when you ask Chat GPT “how do you know that your answers are accurate?”, it will tell you:
“as an AI language model my responses are generated based on patterns and relationships found in the vast amounts of text that I have been trained on. While I strive to provide accurate and helpful information, I cannot guarantee the accuracy or reliability of my responses”
So the onus is on the user to verify any answers generated by Chat GPT. Additionally, Chat GPT has been trained on material that is only as late as 2021. In the legal world, the law can move on significantly within a year, and Chat GPT will have no idea of recent updates in case law or legislation that could significantly alter its answers. It’s likely that this will be true in most other industries too.
At the moment, there are no Terms & Conditions attached to using Chat GPT, so there is no recourse for liability for false information. If you are going to use content generated by Chat GPT in your business, you will need to update your own Terms & Conditions to manage this risk.
Conclusion
If you would like to use Chat GPT in your business, you can protect your business with tightly drafted policies for how your staff can use the site. You may also need to update your Terms and Conditions to deal with any liability arising from content generated by Chat GPT. The risks of using Chat GPT will of course depend on the function to which you put it to use, so if you would like to speak to a lawyer about how to incorporate Chat GPT safely into your business, please get in touch.
