The LLM is based on an optimised and lightweight version of LaMDA, a technological feat Google achieved in 2021. The giant said that Bard, which uses the technology, will become more capable as it gets updated in successive times. However, at its core, the technology is grounded in the search engine’s understanding of quality information.
Although the technology is exciting and emerging, it has shortcomings. It is because the technology has trained itself using information that reflects the biases and stereotypes in the real world. In some cases, therefore, the AI may hallucinate. It means it can provide responses and sound confident, but the answers may be factually inaccurate or misleading.