Bard, Google’s artificial intelligence chatbot, is now available, but only to specific users over the age of 18.
Unlike ChatGPT, which is also very popular, it can get up-to-date information online and has a “Google it” button that lets you search.
It also says where the information came from, such as Wikipedia.
But Google said Bard would have “limitations” and could spread false information and show bias.
This is because it “learns” from data in the real world, which already has these biases. This means that their answers could be based on stereotypes and false information.
What is the purpose of AI chatbots like Bard?
AI chatbots are set up to answer online questions in a way that sounds natural and human.
They can write anything, from computer code to student essays, speeches, and advertising copy.
OpenAI, the company that made ChatGPT, says that within a week of its release in November 2022, more than a million people were using it.
Microsoft has spent billions of dollars on it, and it was added to Bing last month.
It has also said that it will add a version of this technology to its office apps, like Word, Excel, and Powerpoint.
Bard, Google’s version of generative AI that is being released first in the US and UK, has been slower and more cautious than others. Users will have to sign up to try it out.
Bard is a version of Google’s old language model, Lamda, which was never fully released to the public. One of the engineers who worked on it said that its answers were so convincing that he thought it was conscious, bringing it a lot of attention. He was fired because Google didn’t believe what he said.
Google’s senior product director Jack Krawczyk told the BBC that Bard is “an experiment” and that he hopes people will use it as a “launchpad for creativity.”
He showed me how Bard helped him plan his young child’s birthday party.
It came up with a theme that included his child’s love of bunnies and gymnastics, found the address of a place he mentioned, and made suggestions for party games and food.
ChatGPT’s knowledge database only goes as far as 2021, so it can’t answer questions about things like the recent earthquake in Turkey and Syria. On the other hand, Bard can get up-to-date information. For example, it told me about a story on the BBC website a few days ago about how TikTok was banned on government phones in the UK.
It is set up to not respond to offensive prompts, and it has filters that stop it from sharing information that is harmful, illegal, sexually explicit, or could be used to find out who you are. But, Zoubin Ghahramani, vice president of Google Research, said, “These guardrails will sometimes fail, just like any other method.”
I’m right; this is a smart way to start selling something. However, it’s as far as you can get from early big tech’s “move fast and break things” mentality.
When I asked if the company was nervous, Mr. Krawczyk took a moment before saying that the way Bard was launched was “deliberate.”
If Google is scared, it has good reason to be
There is a lot of excitement about this kind of technology, but there are also scary stories about what ChatGPT has been used to do. People also worry that these powerful tools, which are still in their early stages, could be a huge threat to many different kinds of jobs.
Also, and this is important for Google, there is a theory that chatbots could one day take over the very lucrative business of internet search. Why look through pages of search results when you could get one well-written answer? Google needs to keep up.
Mr. Krawczyk and Mr. Ghahramani discussed how some rules and responsibilities come with technology. They also told me about Bard’s huge data centers and how they want to power them with clean energy.
When I asked if students would start doing their homework on Bard instead of ChatGPT, they said Google only lets people over 18 use it. Some teachers are excited about the idea, while others have told their students not to use chatbots to do homework.
Read Also: Chat GPT: Google and Microsoft launches new AI
Google says it will keep a close eye on Bard to ensure it follows its own “AI principles,” which say bias shouldn’t be made or strengthened.
It won’t be able to say what it thinks or have a personality, but it can write like other people can, just like ChatGPT.
Sissie Hsiao and Eli Collins, who helped write the launch blog post for Google, said it helped them write their announcement.