Google Releases ChatGPT-Rival Bard

In an effort to compete in the AI race and preserve its search engine dominance over Microsoft’s ChatGPT-powered Bing, Google has released its long-awaited chatbot Bard.

Google first announced Bard in February, but until now, it was only available to a small group of testers. Beginning this week, the company is opening limited beta access to Bard across the US and the UK to gather more extensive feedback and improve the technology going forward. More capabilities, including coding, multimodal experiences, and languages, are coming.

Interested parties can sign up to test the tool at bard.google.com but must have a Gmail address (Google Workspace emails aren’t accepted). Google is planning to provide access to more countries and languages in the future.

Unlike BingChat, Bard hasn’t been integrated into Google’s search engine; rather, it functions as a separate question box. It’s expected to complement Google Search, but the company has yet to disclose how it plans to leverage the tool’s AI powers without jeopardizing its hefty search engine profits.

From generating useful lists to suggesting writing prompts and explaining complex theories, Bard can do it all. Google’s official blog post describes Bard as a tool that you can use to “boost your productivity, accelerate your ideas and fuel your curiosity.”

Bard runs on Google’s lightweight and optimized version of LaMDA, a large language model (LLM) that operates like a prediction engine. This means that, like similar chatbots, Bard is using large amounts of data to “generate a response by selecting, one word at a time, from words that are likely to come next.”

The most probable response isn’t always the most accurate one, Google warns, advising users to fact-check the results. The company itself failed to fact-check Bard’s faulty results during its first public demo a month ago, which led it to take a more cautious approach in its recent promotion of the chatbot.

Its rival ChatGPT is not immune to mistakes either. Aside from factual errors and fabricated information, the tool has recently leaked payment information, conversation histories, and other sensitive user data.

Google is, however, optimistic that with wider Bard use, the technology behind it will get better at predicting more accurate responses. It also believes that despite Bard’s weaknesses, “there are still incredible benefits to LLMs, like jumpstarting human productivity, creativity, and curiosity.”