Google is attempting to take the lead in the artificial intelligence (AI) race with the announcement of PaLM 2, a “next-generation” large language model (LLM) unveiled at its 2023 annual developer conference (10 May). 

The new general purpose LLM is highly-skilled in software development, maths, natural language generation, and more. 

PaLM 2 is powered by artificial intelligence chips created by Google and is said to be more efficient than other models 

Zoubin Ghahramani, vice president for Google DeepMind, told reporters on Wednesday: “It’s better than our previous state-of-the-art language models.”

Verdict reported on Tuesday that PaLM 2 will include over 100 languages. According to Google, it can also communicate more than 20 programming languages. 

Ghahramani added that around 20 Google products are now using PaLM 2.

Some experts have remained scepitcal about whether Google’s new AI model is enough to make a dent in the highly crowded AI market.

“These advancements showcase the potential impact of AI in transforming industries and improving the quality of life,” Adnan Masood, chief AI architect at UK based digital transformation company, UST, told Verdict.

Masood added: “Having said that, the hallucination grounding, accuracy, and context awareness make a significant difference in the adaptability of such models.

“It is important to recognize that AI advancements are not just about increasing the size of models but also about enhancing their capabilities to better understand and respond to human needs.”

In other parts of the I/O conference, Google showcased the new AI tools implemented into its Maps application. 

The new AI feature will use weather and air quality data to visualise the best route for users to take.  Sundar Pichai, Google CEO, said that it will start to be rolled out in major cities soon. 

GlobalData is the parent company of Verdict and its sister publications.