It’s here: Google’s new AI-powered search engine we’ve been hearing rumors about codenamed Magi.
Is it more “visual, aperitif, personal and human?” Yes.
But for now, you can only access Google’s new Generative Search Experience (SGE) through a Google Labs waiting list, which means you could be waiting weeks before you can play with it directly.
Do not you worry. We’ve taken the first look and deep dive into Google Search’s new generative AI experience.
What the new Google search experience looks like
The interface Google’s new search experience can display an AI-generated answer above search results listings. Google clearly labels the answer as experimental generative AI, which is then followed by an answer to your query.
The answer is bound. Google cites the websites it used to generate the answer. You can click on these sites to learn more. Or you can follow up with an additional question or even click the toggle button in the top right to go deeper.
“You’ll see an AI-driven snapshot of key information to consider, with links to drill down,” Google said.
When you click the expand button to show a deeper answer, you’ll receive additional answers from the generative AI.
Here’s a GIF in action:
Throughout the AI-generated answers, Google gives you websites in these clickable boxes with images, so you can click on the website for more information.
The color of the generative AI answer box will change to “reflect specific trip types and the intent of the query itself,” Google said.
Vertical search with AI. This also works for vertical search experiences, such as Google Shopping results. Google told us that Google SGE can pull 35 billion product listings from the Google Shopping Graph, which has 1.8 billion updates every hour. Generative AI needs to update quickly, almost in real-time, to provide some answers.
Google can give you a good answer on which products to consider when searching for specific types of products, such as [bluetooth speaker for a pool party]:
conversations You can also follow up on your inquiry by adding more details or additional requests in the Request Follow Up box. Google will then generate a follow-up response.
“Context will carry over from one question to the next, helping you continue your exploration more naturally. You’ll also find useful starting points for web content and a range of perspectives you can dig into,” explained Google .
Conversation mode is especially useful for follow-up questions, as well as more complex or evolving information journeys, Google explained.
“It uses artificial intelligence to understand when a person is searching for something that is related to a previous question. It relays the context of previous questions to rephrase the query to better reflect intent,” Google added.
How does this work?
technology Google said this new search experience uses a “variety of LLMs,” including but not limited to MUM and PaLM2.
This search experience was “purposely trained to perform search-specific tasks, including identifying high-quality web results that corroborate the information presented at the output,” Google said.
Where Google won’t give answers. Liz Reid, Google’s vice president of search, told us that Google won’t give you answers to everything you ask. Google is trying to be careful with this new version of Google Search, which will show answers for safer queries.
For example, Google won’t show an answer to a question about giving Tylenol to a child because it’s in the medical space. Google may also not show answers to questions in the financial space.
Sound familiar? Yes, Google is playing it safe in the YMYL (Your Money, Your Life) categories. Google is expanding YMYL to include civic information.
“Just as our rating systems are designed not to unexpectedly shock or offend people with potentially harmful, hateful, or explicit content, SGE is designed not to display such content in its responses,” Google explained.
Google added that they are holding this new search experience “to an even higher standard when it comes to generating answers to certain queries where the quality of the information is vitally important.”
This new search experience “puts even more emphasis on producing informative answers that are corroborated by trusted sources,” Google told us.
When it comes to “data gaps” or “information gaps,” where Google’s systems are less confident in their answers, Google “aims not to generate an AI-based snapshot,” they said.
Also, for explicit or dangerous topics, Google will stay away from generating an answer.
Fluid vs factual. People trust answers given in a more fluid style, Reid told us. But Google prefers to give more factual than fluid answers, because people are more likely to trust a fluid answer.
Hallucinations are a big problem in generative AI, and Google said it’s very sensitive about not giving false or inaccurate information, especially about YMYL topics (eg health, finance).
“Given the trust people place in search, we were intentional about restricting conversation. What this means, for example, is that people may not find conversational mode in SGE to be a free-flowing creative brainstorming partner and instead find it more real with relevant resource metrics,” Google added.
Google’s approach. Google has a five-point approach to generative AI in search:
Information needs: How can Google reduce the number of steps a searcher needs to perform a task or complete a goal, and how can Google make the experience smoother and more fluid?
Quality of information: The information Google responds with must be of high quality and the way the AI responds must be of a high standard. So should Google answer questions related to health or finance?
Security restrictions: Should Google give answers in the first person? Should Google provide smooth answers that users trust to be 100% accurate, when Google might not be able to verify the accuracy of all answers?
Ecosystem: Google wants to provide traffic and credit to the source of the content. Google wants to design an experience that encourages users and the searcher to dig deeper into these sources.
Announcements: Ads can be relevant and provide additional information to the user and how best to show the ads to the user in that experience.
Citations and links
When Google launched Bard, we were all surprised by the lack of citations and links to publishers. It was rare to see links from Google Bard to publisher websites.
However, in Google’s generative search experience, we see a healthier way to connect with publishers and support the ecosystem.
Explicit answers generated in this search experience are not only made up of specific websites, but the websites that make up those answers are also prominently displayed in the answer with a thumbnail image, title, and URL, all of which you can click on the publisher’s website. .
Google, however, will not directly cite or attribute a particular page. Google’s AI model synthesizes information from multiple sources.
In fact, Google looks for fact-checking between sources to create the answers and then display the citations. They usually come from high-quality online sources. Google is using many of the signals Google has had for decades to understand information quality.
Links to publisher sites. Here is a screenshot showing these websites in response:
Change deeper. Then you can click on the top right, this toggle button to drill down to more sources, where the generative AI shows more answers with more sources that you can click on. The arrow in this image points to the switch, directly above the website links:
With search results below. Also, you can continue to scroll down and access the classic search results, in a more “snackable” format. You can see some of the links to the search results, in a more compact format here:
More details
Google talked about it too Principles of AI and emphasized that they take all these AI technologies seriously.
“We’re taking a responsible and deliberate approach to bringing new generative AI capabilities to search,” Google said.
This is not Bard, Bard was designed to show what LLM models can do. This experience is designed specifically for search and works differently, as shown above.
Google has rolled out its search quality raters for some early testing over the next few weeks before rolling it out to the first set of public users. Search Quality Raters will provide feedback both in this pre-launch and ongoing phase to help improve the overall results and experience of this new search approach.
“These qualifications do not directly affect SGE production, but are used to train LLMs and enhance the overall experience,” Google said.
You can waitlist today with the first wave of approvals to try out this new search experience in the coming weeks, more information below.
For more on this topic, check out our accompanying shorter stories:
Why we care
Google’s new generative search experience is a very different search experience than we’ve seen from Google before. But at the same time, it still looks very similar to Google search. Google is also doing a much better job of linking to publishers, with the goal of driving traffic to websites, unlike Bard.
This experience doesn’t replace the Google search you know today; anyone can imagine now when that might happen. We hope Google will listen to feedback and adjust these features before fully rolling them out as the main search experience on Google.com.
[ad_2]
Source link
