Google’s Gemini chatbot has received several updates to enhance its performance and improve user experience. One notable update is the availability of Gemini 1.5 Flash, which was previously only accessible to developers. This low-latency multimodal model is particularly effective in providing quick responses, making it suitable for tasks like answering customer queries. Initially announced during Google I/O, Gemini 1.5 Flash has now been made available to the public. It boasts a large context window of around 1 million tokens, allowing for more complex questions and longer conversations.
To further enhance the capabilities of the free version of Gemini, Google has added the ability to handle file uploads from Google Drive or devices. This feature was previously exclusive to the paid version, Gemini Advanced. By allowing users to upload files, Google is enabling a more comprehensive and efficient interaction with the chatbot.
One common challenge faced by AI models is the issue of hallucinations. In an effort to combat this, Google has implemented a strategy used by other model providers and chatbots: adding related links to prompts that require information. This helps demonstrate that the AI models are not generating information without references. Users can now click on a chip at the end of a paragraph to access additional information on a specific topic, providing a deeper dive into the subject.
In addition to the links, Google will introduce a double-check feature that verifies responses by using Google Search to highlight which statements are corroborated or contradicted on the web. This feature adds another layer of credibility and accuracy to the chatbot’s responses.
It is worth noting that Google is not the only company incorporating links for attribution in their chatbots. ChatGPT and Perplexity also include citations and links to relevant websites. However, a report from Nieman Labs raised concerns about hallucinated links, where chatbots attached links to non-existent or unrelated news stories.
Overall, Google’s updates to the Gemini chatbot, including the availability of Gemini 1.5 Flash and the addition of source links, aim to provide users with more accurate and reliable information. By addressing challenges such as hallucinations and enabling file uploads, Google is continuously improving the functionality and user experience of its chatbot. These updates demonstrate Google’s commitment to enhancing the capabilities of AI models and delivering high-quality conversational AI to users worldwide.