News about chatbots including Open AI’s ChatGPT and Google’s Bard has been dominating headlines and social platforms, with users expressing fears from the seemingly sentient responses of this new tech to the potential legal pitfalls its use could pose.

ChatGPT, Bing AI and Google Bard are AI-powered chatbots that automatically generate human-like responses to queries. They rely on scraping content from a variety of sources and are programmed to respond in a conversational manner.

While ChatGPT can successfully condense long, complicated topics into a several paragraph summary with reasonable accuracy, these technologies are far from perfect, and their use could leave companies open to liability on a number of fronts.

Take steps to limit liability

As with any new technology, the legal dangers of these evolving AI chatbots have yet to be determined. No laws or regulations on this type of artificial intelligence currently exist, and experts fear widespread and unchecked business use of chatbots could lead to claims alleging bias, copyright, libel or slander.¹

One of the biggest concerns involves the ability to check for accuracy. OpenAI has stated that ChatGPT does not use copyrighted materials, but with no safeguards to check whether the bot has regurgitated copyrighted material, and no citations to check accuracy, users that pass off ChatGPT-produced content as their own could be opening themselves up to plagiarism and E&O claims.

There’s also the risk of releasing biased content that could lead to claims of discrimination and cause severe reputational damage.

Since the technology is so new, no one knows how insurance may respond. However, some insurance coverages that could be tested in the future by this technology include cyber, errors and omissions and general liability.

For now, companies allowing the use of AI bots like ChatGPT should take these steps to protect themselves from the unknown:

• Classify its use as a trial. ChatGPT itself is in a free trial phase. Businesses may want to allow workers to try out the technology, but need to be clear that it is temporary until it can be further evaluated.
• Fact-check any business use of the technology. ChatGPT can create content at an astounding speed, but inaccuracies exist. OpenAI has acknowledged that the software may create content that includes bias. Before relying on any bot-produced documents, make sure to have a human edit and fact check.
• Create policies around usage. If your company plans to use ChatGPT or another AI bot as part of its business processes, ensure that polices are crafted to define the limits of usage and the steps taken to verify accuracy.

Contact HUB International’s ProEx insurance experts to learn more about how new technologies can affect your overall risk.

¹Reuters, “As ChatGPT’s Popularity Explodes, U.S. Lawmakers Take an Interest,” February 13, 2023.