Bold Claims with its Intelligence! Einstein Copilot Introduced
Salesforce introduced its new chatbot named Einstein Copilot. This artificial intelligence chatbot minimizes the possibility of misunderstanding.
Salesforce has released its new chatbot, Einstein Copilot, designed for businesses, to the general public. According to Salesforce executives, Einstein Copilot carries less risk of hallucination (i.e., generating incorrect or meaningless information) compared to other AI chatbots. Thus, it aims to address the issues commonly encountered by other AI chatbots.
Salesforce's new artificial intelligence chatbot, Einstein Copilot, has been introduced.
Patrick Stokes, Salesforce’s Vice President of Product Marketing, points out that Einstein Copilot is different from other AI chatbots. This difference stems from the bot’s use of a company’s own data. In other words, the bot scans the company’s internal data before being sent to large language models (LLMs), ensuring it works with accurate information. This reduces the risk of producing incorrect information.
In addition, Einstein Copilot includes a security layer to protect the data shared with LLMs. This aims to prevent misuse of company data by chatbots. Furthermore, the bot collects real-time customer feedback to identify weaknesses in the system and includes a feature to detect hallucinations.
Stokes acknowledges that hallucinations cannot be completely prevented. However, he states that they are working to develop transparent and secure technologies to keep them to the lowest possible level. Salesforce’s CMO, Ariel Kelmen, notes that large language models (LLMs) are inherently designed to imagine, hence the possibility of hallucinations.
Therefore, Salesforce aims to minimize the risk of hallucination compared to other AI chatbots with Einstein Copilot. However, they also acknowledge that it is not possible to completely eliminate hallucinations in the world of artificial intelligence. This is one of the significant challenges the industry faces regarding the reliability and accuracy of AI.
