Google Launches Experimental Chatbot Bard to Take on Microsoft’s ChatGPT

Google has taken the AI industry by storm with the launch of its new experimental chatbot, Bard. Aimed at countering the popularity of Microsoft’s ChatGPT tool, Bard is designed to combine the breadth of the world’s knowledge with the power and creativity of Google’s large language models.

The announcement of Bard comes just two weeks after Microsoft revealed its investment in OpenAI, the maker of ChatGPT and other AI tools that can write readable text and generate new images. ChatGPT, which was released in November 2020, has caused a sensation in the AI community for its ability to write essays, poems, and even programming code on demand within seconds.

Google CEO Sundar Pichai wrote in a blog post that Bard seeks to provide fresh, high-quality responses by drawing on information from the web. Unlike ChatGPT, Bard is based on LaMDA, Google’s Language Model for Dialogue Applications system, and has been in development for several years.

Pichai emphasized that the responses generated by Bard will meet a high bar for quality, safety, and real-world information. This is in response to the fears of cheating and entire professions becoming obsolete that have arisen with the widespread use of AI tools like ChatGPT. The CEO also stated that Bard would source its responses from a limited version of its base language model to reduce computing power and reach a wider audience.

The launch of Bard is being seen as a direct challenge to Microsoft, which is integrating ChatGPT into its Teams platform, with expectations that it will eventually adapt the app to its Office suite and Bing search engine. Independent tech analyst Rob Enderle says that generative AI has the potential to change the competitive dynamic for search and information and warns that Google’s dominance in search engines could be threatened if it fails to adapt.

Google has also hinted that users will soon see AI-powered features in its search engine, with new-style responses designed to “distill complex information and multiple perspectives into easy-to-digest formats.” Thierry Poibeau of the CNRS research center in Paris says that search engines powered by generative AI will give structured answers to questions, rather than just links, but warns that chatbots like ChatGPT can also provide incorrect answers.

Before the emergence of ChatGPT, Google had been hesitant to launch its own language-based AI, wary of the reputational risk of releasing technology that was not ready. Researchers using the same language models as Bard or ChatGPT have demonstrated the technology’s ability to spread misinformation or nonsense on a massive scale. Facebook-owned Meta had to take down its own large language model, Galactica, just three days after its release due to its biased and incorrect results being shared on social media.