After months of delay, Bard finally launches in the European Union. Google’s AI software will try to challenge ChatGPT’s dominance.
Months after ChatGPT was released, tech giant Google replied with its own version: Bard. The AI software takes the nickname of playwright genius William Shakespeare.
The software launched on Thursday in Europe and Brazil, but its original release date in the United States was March 21st. Launch in the European Union was delayed, as it often happens with tech softwares, by privacy violation issues.
The Irish Data Protection Commission pointed at Google’s lack of explanation on how Bard would protect EU citizen privacy. This is not the first time Google has to change a software’s privacy settings for the European market.
Even ChatGPT was briefly banned in Italy and other EU countries for similar concerns. When OpenAI, ChatGPT’s parent company, solved the issues the software became available again.
OpenAI has been acquired by Microsoft in a $10 billion deal, starting competition in the nascent artificial intelligence market with Google. Microsoft has integrated AI capabilities into its search engine Bing, creating a chatbot connected to the internet.
Following the quick AI market craze, the public seems to be losing interest in this new technology, with ChatGPT witnessing a decline in users for the first time in June. And it all might be related to one problem.
The problem with artificial intelligence
Despite their great technological advancement, neither Bard nor ChatGPT nor Bing Chatbot can solve AI’s main issue: accuracy. None of these chatbots can give accurate answers: indeed, the problem is how confident they sound in their completely wrong statements.
For the moment, AI chatbots are simply predictive algorithms that try to understand which word comes next in the sequence they’re writing. Chatbots do not understand language, and indeed do not understand questions either. They only use the entire array of human text on the internet to predict how to make a grammarly correct phrase.
Despite having access to the world’s most reliable search engine, Bard cannot yet give fully accurate answers. In the very statement at the beginning of Bard’s "Privacy terms" statement Google writes "Bard is an experimental technology and can provide inaccurate or inappropriateinformation thatt does not represent Google’s opinions."
Google has developed other types of artificial intelligence using Large Language Models (the technology of Bard and ChatGPT). Most recently, Google developed a "medical" AI capable of giving sound advice and research to healthcare professionals.
The new software, called "Med-PaLM 2" is currently under testing at the Mayo Clinic research center. It has, again, a significant accuracy problem that does not allow its use by real medical professionals.
