Google launches Bard and other reasons AI ain’t going away

21 Mar 2023

Image generated by AI: © Pascal M/Stock.adobe.com

Bard is now available in the US and UK, Microsoft has added Dall-E to Bing and Adobe has its own image generator. Here’s what’s happening in AI.

After months of anticipation, Google has finally launched its generative AI chatbot Bard in the US and the UK.

Described by Google as “an early experiment that lets you collaborate with generative AI”, Bard is the search giant’s response to OpenAI’s ChatGPT. It can perform text-based tasks ranging from suggesting tips on reaching annual reading goals to explaining quantum physics in simple terms.

That chatbot, only available to over 18s for now, is powered by a large language model that Google says is a lightweight and optimised version of LaMDA – its system for building chatbots. Over time, Google will update Bard with newer, more capable models.

However, the company warned that Bard will not be without its faults in this early stage.

“Because [large language models] learn from a wide range of information that reflects real-world biases and stereotypes, those sometimes show up in their outputs,” Google executives Sissie Hsiao and Eli Collins wrote in a blogpost today (21 March).

“And they can provide inaccurate, misleading or false information while presenting it confidently.”

They gave the example of when Bard was asked to share a couple suggestions for easy indoor plants and got some things wrong, such as the scientific name of the ZZ plant.

“And so, when using Bard, you’ll often get the choice of a few different drafts of its response so you can pick the best starting point for you. You can continue to collaborate with Bard from there, asking follow-up questions,” they wrote.

“And if you want to see an alternative, you can always have Bard try again.”

Microsoft brings Dall-E to Bing

Meanwhile, Microsoft has just rolled out an AI image creation feature on Bing that will use OpenAI’s Dall-E to create text-prompted images.

Users of the feature, called Bing Image Creator, will be able to ask Bing to create images based on descriptions they provide, such as an astronaut walking through a galaxy of sunflowers or tips for redecorating a living room with mid-century modern furniture.

It will also be available on the Microsoft Edge browser through the search bar.

Earlier this month, the software giant, which has invested billions in OpenAI, said that it is expanding its suite of business applications tools to include new AI capabilities. Last week, Microsoft announced it is bringing AI to popular apps such as Word, PowerPoint and Excel.

Adobe Firefly takes flight

Photoshop creator Adobe is also banking on artificial intelligence for future products, with the launch of its own “family of creative generative AI models” called Adobe Firefly.

The Silicon Valley-based software giant has already released two of these tools: an AI-powered image generator that works on descriptive prompts and a stylised text generator that resembled WordArt.

In a blogpost today, Adobe executive Dana Rao stressed on how ethics is at the heart of the company’s explorations of AI’s potential.

“Adobe’s first model in our Firefly family of creative generative AI models is trained on Adobe Stock images, openly licensed content, and public domain content where copyright has expired,” he explained.

“Training on curated, diverse datasets inherently gives your model a competitive edge when it comes to producing commercially safe and ethical results.”

Now in beta mode, Firefly will be integrated directly into Creative Cloud, Document Cloud, Experience Cloud and Adobe Express workflows.

OpenAI slips up

And while these software giants were releasing their respective AI features and innovations, OpenAI – the company that arguably triggered the recent AI race – was busy fixing a major bug that temporarily exposed AI chat histories to other users.

The company had to temporarily shut down ChatGPT after it emerged that some users were being able to see AI chat titles of conversations that weren’t their own. An OpenAI spokesperson told Bloomberg that the bug did not share full transcripts of conversations but only brief titles.

“We are not able to delete specific prompts from your history,” reads an FAQ on OpenAI’s website. “Please don’t share any sensitive information in your conversations.”

10 things you need to know direct to your inbox every weekday. Sign up for the Daily Brief, Silicon Republic’s digest of essential sci-tech news.

Vish Gain is a journalist with Silicon Republic

editorial@siliconrepublic.com