There’s been a lot of chatter around artificial intelligence, especially after OpenAI debuted its revolutionary ChatGPT service, which Microsoft integratedinto Microsoft 365’s Co-Pilotand Bing. Meanwhile, Google has been a self-professed “AI-first” company since announcing a shift in focus at I/O 2017. The company unveiled a newAI-powered chatbot called Bardin March 2023. But it can be hard to understand what these new tools do.
Google’s suite of tools, from writing emails to searching online, uses AI to understand colloquial language and power tools likeGoogle LensandGoogle Assistant. You might wonder how Bard is different. The key is in Bard’s conversationality and ability to answer questions. Still, there’s more to it, so let’s dive in.
What is Bard, and where did it come from?
In the simplest terms, Bard is a generative artificial intelligence (AI) interface that uses Google’s large language models. Other popular generative AIs are ChatGPT and DALLE-E. Generative AIs can make video, audio, and imagery. As an AI chatbot, Bard focuses on creating text that answers your questions naturally and conversationally.
Another umbrella term that describes Bard is a large language model (LLM). This neural network has been fed vast amounts of text to teach it to process natural language. The training data is one limiting factor in generative AI development. There are many mathematical solutions to this limit, and most require a large amount of server space.
Google’s approach is more streamlined than OpenAI’s ChatGPT LLMs. Google focuses on accurate, human-customized training. The AI’s machine learning breaks down languages into data it can understand and then imitate. It continually corrects itself to get a closer approximation to content like speech or images. Some of Google’s AI work helps businesses create their own AIs with such technology. Google Bard is consumer-focused, so anyone can try it out right now.
Considering the timing, it may seem like Bard was rushed out the door to compete with the release of ChatGPT 3 and 4. Google laid the groundwork for ChatGPT and other LLMs by making its Transformer deep learning model available to the public in 2017. Bard is powered by Google’s Language Model for Dialogue Applications (LaMDA), which released in 2021. OpenAI’s new tool shares a common lineage with Google’s, but Bard has been years in the making.
Bard is based on LaMDA, a conversational AI model Google debuted in 2021.
Meanwhile, Bard has become a top priority for Google. The company recentlyrestructured its Assistant teamto bring more AI specialists into the fold on Bard’s development. It’s steadily addedfeaturesto the chatbot since its debut. So much so that itmaintains a changelog, and it’s safe to say Bard is here to stay.
What does Google Bard’s name mean?
Bard gets its name from the word “poet,” as in the Bard of Avalon by William Shakespeare, in reference to its linguistic capabilities. The name also has geeky roots, which fits Google’s naming conventions. Bards are a type of playable character in Dungeons and Dragons. They inspired the Apple II game called The Bard’s Tale, which spawned a series of games that includes a remastered 2018 edition of the original.
How does Bard work?
Google wants Bard to supplement the Knowledge Graph Cards you see in Search when making queries that have a simple answer. While a Knowledge Graph Card can supply a word’s definition or an overview of a person or place, Bard’s responses address NORA questions (searches with No One Right Answer), as Google calls them.
To do this, Bard use various LLMs to understand your question and its context. Since Google’s models use large datasets that contain dialogue, thet understand nuance and colloquialisms that search engines struggle with. After that, Bard draws on information it finds on the web to form an answer, which is made into the type of conversational reply you might expect from a real person (again, thanks to LaMDA). Bard’s goal, as well as all AI chatbots, is to provide high-quality responses.
Google wants you to use this tool to further your understanding of topics and make decisions. Duringa demonstration in Paris, the company asked the AI chatbot to help decide which car to buy and then asked follow-up questions about the advantages of electric vehicles. Such prowess could negate the need to click search results, but Google is careful to maintain its relationships with websites and content creators. Senior Vice President Prabhakar Raghavan had this to say on the matter:
As we scale these new, generative AI features in our search results, we continue to prioritize the process that will allow us to send valuable traffic to a wide range of creators and support a healthy, open web.
Google may be using some of its AI tech behind the scenes for Google Search RankBrain or Google Assistant. In its current form, Bard is a standalone utility rather than a connected function of Google Search. You can open a chat with the AI, and the things you say during a session are remembered, allowing you to use the context of earlier questions to get refined answers from Bard. You can also press the chatbot for clarity when the information it gives seems off, and it corrects itself and apologizes.
Google Bard and Google Gemini
Google announced Gemini, its latest multimodal LLM, in late 2023. Gemini is primarily a business-facing suite of products expected to launch in 2024, starting with the early release of Gemini Pro. It’s a professional LLM, like other Google projects such as PaLM 2, but is highly multimodal. It is cross-trained in all kinds of content, from visuals, memes, and emojis to text, diagrams, and speech. Gemini is one of the most interconnected AIs, which helps it synthesize data from many sources. In one of Google’s famous examples, it recognized a drawing of a blue duck and theorized that blue ducks don’t exist in the wild, only to correct itself when shown a blue rubber ducky.
Along with the release of Gemini, Google announced that Gemini Pro is incorporated with Google Bard, which benefits — as much as possible for a conversation-focused AI — from Gemini Pro’s capabilities. Gemini Pro beat ChatGPT 3 in terms of accuracy, but ChatGPT 4 remains slightly ahead.
What are the ethical implications of Bard and generative AI?
Bard, ChatGPT-4, and other AI chatbots may provide quick answers, but they may not be correct. These platforms provide answers that are influenced bybiases in training datasets, conflicting or outdated information, or, in the case ofCNET’s Wordsmith, facts the AI makes up.
Google has a lot of experience sorting fact from fiction and determining when to surface safe opinions. That is, opinions and theories that are mostly accepted as fact and surface the most accurate information available at the time.Google’s Featured Snippets, for example, answers queries bycollecting data from multiple online sources, determines a midrange using machine learning models toquantify sentimentand other variables, filters out answers that fall outside the midrange, and presents a response that precedes its search responses.
After more than a decade at Google, Dr. Geoffrey Hinton, known in some circles as The Godfather of AI, resigned from his position andvoiced concernsabout the future of generative AI. The researcher believes artificial intelligence models are advancing rapidly, will be hard to regulate, and could become a formidable tool for bad actors.
“Look at how it was five years ago and how it is now,” said Hinton. “Take the difference and propagate it forward. That’s scary.” His unease is mainly due to an industry-wide push to win users, potentially at the expense of due diligence. Google says it “remains committed to a responsible approach to AI.” Hinton seems to feel that while Google had been a “proper steward” of the technology until recently, the rise of ChatGPT forced the company to push forward with Bard in haste.
Meanwhile, more than a thousand industry leaderscalled for a six-month moratoriumon developing artificial intelligence models in an open letter warning of “profound risks to society and humanity.” Mirroring Hinton’s sentiment, the letter alleges that companies like Google, Microsoft, and OpenAI are “locked in an out-of-control race to develop and deploy ever more powerful digital minds that no one, not even their creators, can understand, predict, or reliably control.”
In aninterview with 60 Minutes, Google CEO Sundar Pichai tried to ease some of these concerns by reiterating Google’s commitment to responsible AI development. However, he admitted, “There is an aspect of this which we call … a ‘black box.’ You know, you don’t fully understand.”
Problems continue on this front.Google is being suedfor allegedly stealing copyrighted data to use for training materials, and Google is suing others for imitating Bard services to scam people. Ongoing concerns remain about generative AI’s impact on society.
Google made it clear that its AI technologies are aimed at businesses. Those services could take jobs from humans (something that Googlepushes against in the Bard FAQs) or make it easier for ads to exploit the unwary, among other unethical tasks. Conversations on these subjects are ongoing and murky.
How can I use Bard?
After a trial period, Google began rolling out Bard to many users,so you can visit the site, sign the waiver, and try out Bard.
Chatbots like this take a little practice to get used to, especially if you want them to generate more complex content instead of general answers. you may learnhow to use Google Bard to find the information you’re looking for. Bard can do things like:
Bard can get things wrong, and its language can be stilted, rough, or repetitive, so don’t rely on it. Proofread before incorporating its content into anything.
While there isn’t a dedicated mobile app for Bard, it’s easy to access the tool through your smartphone’s browser app. If you use a Google Pixel phone,Google is working on a Bard widgetthat will provide faster access to the tool. There’s no word on when this might roll out widely, but we expect to see it on some Android phones shortly after it’s made available toGoogle Pixel phones.
Make Bard work for you
Google Bard’s AI technology isn’t only at work in Bard. The LLMs that Google uses to fuel Bard work in various forms to help youwrite emails in Gmail, fill out Google Docs, provide real-time information in Google Maps, and more. If you use Google Services, you’re using part of Bard.
