Google’s latest innovation, the Search Generative Experience (SGE), revolutionizes the search process by incorporating AI-generated responses directly above search results. These responses are meticulously curated, complete with citations to the sources that contributed to formulating the answer.
The SGE, previously known by its codename “Magi,” represents a substantial leap forward in search technology. It is designed to offer a more visual, personalized, and human-centered approach to information retrieval.
Accessing this cutting-edge search experience is currently limited to those who join the waitlist in Google Labs. Due to the high demand, there may be a waiting period of several weeks before users can explore it firsthand.
Key Features of the Google SGE
User Interface. In the updated Google search experience, you’ll encounter AI-generated responses preceding the search results. Google explicitly indicates these as experimental applications of Generative AI. Each response is accompanied by references to the consulted websites, allowing further exploration.
Expanded Responses. Clicking the expand button provides additional answers generated by the AI model, each accompanied by clickable boxes containing website links and images for further information.
Vertical Search with AI. Specialized search experiences, such as Google Shopping results, benefit from the SGE. It aggregates an impressive 35 billion product listings from the Google Shopping Graph, undergoing constant updates to provide accurate and timely information.
Conversational Mode. Users can extend their inquiries in the “Ask a follow-up” box, allowing Google to generate follow-up responses. Contextual information is maintained for seamless exploration.
The Technology Behind the SGE
Google employs a diverse array of Language Learning Models (LLMs), including models like MUM and PaLM2, to power this fresh search experience. It is intentionally trained to excel in Search-related tasks and ensure accurate results.
Safety and Trust
Google places a high premium on delivering reliable and trustworthy information. The SGE undergoes stringent quality checks, especially in critical areas like Your Money, Your Life (YMYL) topics like health and finance. The goal is to provide informative responses supported by credible sources.
Google’s Five-Point Strategy for Generative AI in Search
- Information Needs – Streamlining the search process for users to achieve tasks with fewer steps.
- Information Quality – Ensuring high-quality information is provided in responses, particularly for health and finance inquiries.
- Safety Constraints – Offering answers in a natural and trustworthy manner, even when independent verification is impossible.
- Ecosystem – Directing traffic and acknowledging the content source to motivate further exploration.
- Ads – Ensuring pertinent ads offer supplementary information to users in a non-intrusive manner.
Citations and Links
Unlike the previous Bard model, SGE adopts a robust approach to linking with publishers, prominently showcasing specific websites within the provided solution. Google relies on signals it has used for years to evaluate information quality.
Google emphasizes its commitment to responsible AI use, adopting a conscientious approach in introducing new generative AI capabilities to Search. Preliminary testing by search quality raters will enhance the overall user experience before rolling it out to the public.
Google’s new Search Generative Experience represents a significant advancement in search technology. It offers a more personalized and informative search experience, focusing on delivering accurate, trustworthy responses. While it’s currently accessible through a waitlist, its potential to reshape how we interact with search results is undeniable. Stay tuned for further updates as Google refines this innovative approach.