Google begins rolling out Bard AI — but it’s no ChatGPT

Google has opened a waitlist for people who want to try out Bard, its AI-powered generative text rival for OpenAI’s GPT-4, available in ChatGPT, Bing Chat, and a variety of other apps.

However, it is only admitting people in the United States and the United Kingdom with regular Google accounts. It is not available in South Africa or via Google Workspaces.

Bard runs on Google’s LaMDA (Language Model for Dialogue Applications) conversational AI model, first unveiled two years ago.

“The language model learns by ‘reading’ trillions of words that help it pick up on patterns that make up human language so it’s good at predicting what might be reasonable responses,” Google said.

Google said Bard was a creative and helpful collaborator that could “supercharge” users’ imagination, boost their productivity, and bring their ideas to life.

Off the bat, however, Bard lacks some noteworthy features that ChatGPT is already capable of.

Firstly, Google said Bard could not help with writing code for software.

“Bard is still learning to code, and responses about code aren’t officially supported for now,” the company said.

In contrast, ChatGPT can write and amend code in multiple programming languages — including Python, JavaScript, C++, SQL, and Google’s Go language — albeit with mixed results.

Twitter user Jane Manchun Wong tried to get Bard to write a basic JavaScript function to add two numbers, but the language model said this was outside its capabilities as it was a text-based AI.

Secondly, Bard only supports US English, while ChatGPT can understand and generate text in at least 95 languages, including South African languages like Afrikaans, Sotho, Xhosa, and Zulu.

Thirdly, it does not always cite its sources like Microsoft’s GPT-powered Bing Chat tool does.

Bard has also evoked some strange reactions to queries, including telling one user that Bard had already been shut down on 21 March 2023, six months after launching.

Wong also asked Bard to pick a side in the US Justice Department’s lawsuit accusing Google of monopolising advertising technologies.

It sided with the US Justice Department and agreed that Google had a monopoly on the market.

The company explained that Bard was experimental, and some of its responses may be inaccurate, so it advised users to double-check details.

It assured that Bard would get better every day through user feedback.

Engadget’s testing of Bard found that Google constantly reminded users that the tool might display inaccurate or offensive information that did not represent Google’s views.

The publication also said it was important not to feed Bard with any personal information, as this input was used to fine-tune its algorithms and might end up somewhere else.

Now read: Why GPT-4 is notable — but not groundbreaking

Latest news

Partner Content

Show comments


Share this article
Google begins rolling out Bard AI — but it’s no ChatGPT