The company says its artificial intelligence will be developed with care as it is launched in 180 countries.
Google is rolling out its AI chatbot Bard throughout 180 countries worldwide while promising that it will be developing this technology ethically.
Bard had initially rolled out exclusively in the US and UK, but it has announced a broader rollout.
Last week at the annual I/O conference, Google made an announcement that it would be opening up its AI chatbot to many more countries worldwide. While this is big news, it has been met with a spectrum of responses, including those who are highly enthusiastic about it, and others who feel outright threatened and angered or saddened by it.
Bard is built on the company’s PaLM2 large language model in a way comparable to how ChatGPT Is built on the GPT from OpenAI. It is able to write code, translate languages, analyze images and provide information.
The Bard AI chatbot will experience more advancements over time following the rollout.
For instance, at the event, it was also announced that Bard will soon be providing visual responses as well as the text-based responses. Moreover, users of the Lens application will be able to upload images for analysis by Bard. The example Google gave to illustrate this feature used an image of two dogs and the instruction to “write a funny caption for these two”. From there, Bard will have the capability to create responses while also having identified which breed the dogs are.
Bard will be including annotations on information it has sourced from somewhere else and will offer a link to that source. This will help to overcome the issue in which AI chatbots create information or a sourced text that it claims to be true, known as AI hallucinations.
Right now, Google is working on making Bard available in over 40 languages. At the moment, the only options available are English, Japanese and Korean. Another reason the technology has launched slowly in additional languages is that the company’s early research has shown that systems built on PaLM2 “continue to produce toxic language harms.”