Google unveiled its “state-of-the-art language model” PaLM 2 at its I/O 2023 developer conference on Wednesday. Google said it introduced more than 25 new products and features powered by PaLM 2 at I/O 2023. Google also opened up access to the PaLM 2-powered Bard chatbot in over 180 countries and territories, including India. Google claims to have added advanced privacy controls into PaLM 2 that allow users to have full control over their personal information.
Previously, users in the US and UK had to go through a waitlist to access Bard.
Announcing the launch of the PaLM 2 language model in a blog post, Google said, “PaLM 2 shows us the impact of highly capable models of various sizes and speeds — and that versatile AI models reap real benefits for everyone”
What is PaLM 2?
PaLM 2 is a language model that Google is deploying to bring AI capabilities to all of its products, including Gmail, Google Docs and Bard. Similar to other language models like GPT-4, PaLM 2 is capable of powering AI-based chatbots. It is also adept at writing code, analysing images and translating. In fact, PaLM 2’s multilingual capabilities will be used to make Bard available in over 40 languages.
According to Google, PaLM 2 has been trained on multilingual texts from over 100 languages, and the language model has also passed advanced language proficiency exams at the ‘mastery’ level. PaLM 2 has been trained on publicly available source code datasets to make it proficient in over 20 programming languages, including Java, Python, Ruby, C, etc.
Where is PaLM 2 being used:
A number of upcoming and current Google products will be powered by the PaLM 2 language model. Most notably, users can get a taste of PaLM 2’s capabilities by accessing the Bard chatbot. PaLM 2 is also being used to add artificial intelligence capabilities to Google Docs, Gmail and Google Sheets. Google is also using PaLM 2 in Sec-PaLM for cybersecurity analysis and in Med-PaLM 2 to answer questions and summarise insights from a variety of dense medical texts.
What’s next?
Google has also announced that it is already working on the next-generation language model, called Gemini. While still in training, Google says Gemini is already demonstrating multimodal capabilities not seen in its previous models.
“Gemini is highly efficient at integrating tools and APIs, and is designed to enable future innovations such as memory and planning.” the search giant added
Google has been struggling to find its footing in the artificial intelligence landscape since the launch of ChatGPT in November last year. The launch of GPT-4 with multimodal capabilities has only compounded Google’s problems, with OpenAI’s language model being used to power Microsoft’s new Bing search. Microsoft and Google are vying for supremacy in the field of generative AI-powered chatbots.
Download The Mint News App to get Daily Market Updates & Live Business News.
