GenAI is highly inaccurate for business use -- and getting more opaque.

Generative AI platforms, such as ChatGPT, are highly inaccurate and becoming less transparent when connected to corporate databases, according to two studies. The Stanford University study found that as these platforms ingest more data and grow in size, it becomes harder to track the origin of the data they use. This lack of transparency makes it difficult for businesses to build applications using commercial genAI models and for academics to rely on them for research. Additionally, the overall accuracy of these models has been questioned, with one study showing that they return accurate responses to basic business queries only 22% of the time. The inaccuracy of these models erodes trust and can have significant consequences for businesses. To address these issues, companies need to invest in strong data foundations, such as Knowledge Graphs, and ensure oversight and governance to protect sensitive information. [Extracted from the article]