Google’s Bard – The Controversial AI Chatbot
Google already employs AI in everything from search to dictation and call screening. Its newest addition is a chatbot called Bard, which aims to answer questions in a conversational manner — and it’s already raising eyebrows. In fact, Bard Google is so controversial that former and current employees accuse the company of ignoring internal safety and ethics rules in a rush to beat rival products like Microsoft’s Bing Chatbot and OpenAI’s DALL-E.
In February, one Google employee posted in an internal message group that the tool was “worse than useless” and urged executives not to launch it. Others agreed that the software was prone to erroneous or even egregious responses on simple factual queries. Google’s Jen Gennai, the company’s AI principles ops and governance lead, overruled that assessment and approved Bard AI release, according to sources quoted in a Bloomberg story published on Wednesday.
The problem, as explained by those employees interviewed, is that Bard AI is a generative AI. That’s the generic term for AI models that create new content, including video, audio, and imagery. But the type of generative AI that powers ChatGPT and DALL-E is also capable of producing text, and in some cases, that can be dangerous. For example, if you ask the Google Bard AI chatbot how to land a plane, it may confidently respond with instructions likely to result in a crash.
However, unlike some generative AIs, Google AI Bard is plugged directly into the web, which gives it access to fresh information and allows it to correct itself when it’s wrong. Bard AI use is also simplified with a button that lets you send feedback on a response and the option to use an existing question or create a new one.
Despite its many problems, the bot is still in wide public use. It’s available to anyone in 180 countries who signs up for a Bard account on the company’s website. Google says it will add support for more languages and features but didn’t give a specific timeline for those releases.
At I/O 2023 on May 10, Google chief executive officer Sundar Pichai took the stage, but it was General Manager of Google Assistant Sissie Hsiao who stole the show by announcing that Bard would now be available without a waitlist in more than 180 countries worldwide. Those users will be able to communicate with the bot in English, Japanese, and Korean and will be able to use a button that seems like it’s similar to the “Tools” button on ChatGPT that allows you to plug in additional language models.
The company also plans to make creating and sharing a custom bard easier. Also, the app has now been made available on iOS and Android. This is a big update for the AI that drew criticism from its first day in public because it was giving out information that could potentially cause harm. Google recommends that people always double-check the responses of Bard AI Google and carefully test and review them for errors, vulnerabilities, and risks.