The availability of ChatGPT is quickly reshaping how organizations solve problems, as many users take advantage of its considerable capabilities to search for important information or generate code to address specific needs. However, while these features can be big time savers in the short term, organizations need to be careful to enact policies around its use that carefully consider what is being provided to ChatGPT as inputs, to ensure sensitive information doesn’t accidentally make its way into the public domain, according to an article in VentureBeat.
Already, multiple organizations have found critical information exposed, with Samsung employees providing ChatGPT with source code used to resolve programming errors, and Amazon, Walmart, and JPMorgan also finding instances of employees providing the tool with sensitive information. Concerns also persist with respect to skilled users finding ways to pull personal data from ChatGPT and other tools, whether through security breaches or data extraction attacks.
While it’s likely companies will continue to look toward ChatGPT and its competitors, they’re also becoming increasingly aware of the need to heavily limit how it can be used.
Source:
https://venturebeat.com/ai/mind-your-language-risks-using-ai-powered-chatbots-chatgpt/
