Meta is reportedly creating a custom search engine for its AI chatbot.
Meta Platforms Launches Custom Search Engine for AI Chatbot
Meta Platforms Inc. is set to introduce a new custom search engine that will be integrated into its AI chatbot, as reported by The Information. This move aims to enhance user access to current event information directly through the chatbot, thereby reducing reliance on major providers like Google and Microsoft.
Overview of Meta AI Chatbot
Features of Meta AI
- Launch: The Meta AI chatbot, similar to ChatGPT, was launched the previous year.
- Accessibility: Users can access it across various platforms, including Facebook, Instagram, WhatsApp, and Messenger, as well as through a dedicated website.
- Technology Powering the Chatbot: The chatbot operates using Llama, a series of large language models that Meta started developing early last year. The most advanced model, Llama 3.1 405B, was introduced in July and shown to outperform OpenAI’s GPT-4o in several benchmark tests.
Current Dependence on Third-Party Services
Presently, Meta AI relies on technologies from Google and Microsoft for retrieving timely information. However, the reliance on these platforms presents a risk—if either company changes its services, Meta’s chatbot could lose access to essential data.
Development of Meta’s Custom Search Engine
Meta’s initiative to create a custom search engine started over eight months ago, spearheaded by Xueyuan Su, a senior engineering manager. This search engine’s development will likely allow Meta to handle information sourcing internally, leading to several potential advantages:
Benefits of an In-House Search Engine
- Reduced Costs: Currently, Google and Microsoft provide access to their search engines through paid APIs. By developing their own search engine, Meta can save on these costs.
- Enhanced Customization: An in-house engine can be tailored to meet the specific needs of the Meta AI chatbot, offering features that rival standard commercial solutions.
- Mitigated Risk: By owning the search engine, Meta won’t face issues stemming from third-party changes, ensuring continued access to necessary data.
Speculations on Search Engine Functionality
While it remains unclear how Meta’s search engine will function, two possibilities arise:
- Standard Web Links: The engine could return a list of webpage links based on user queries.
- AI-Generated Responses: Alternatively, it might produce AI-driven answers that directly respond to user inquiries, similar to the capabilities of OpenAI’s SearchGPT, which launched in July.
Competitive Landscape
Meta’s move to develop its own search capability follows a series of strategies designed to enhance the competitiveness of Meta AI in comparison to ChatGPT. Recently, the company entered into a partnership with Reuters to provide news content within its chatbot.
Staying competitive in this landscape is essential, especially given that OpenAI has secured several content licensing deals with various publishers. Additionally, in September, Meta AI gained the ability to answer questions regarding images uploaded by users, a feature that has been available on ChatGPT since 2023.
Current User Base and Revenue Insights
As of August, Meta AI boasts around 185 million weekly users, while ChatGPT has approximately 250 million. OpenAI disclosed that a significant portion (75%) of its revenue, projected to reach $3.7 billion by year-end, comes from consumer subscriptions. This financial model may motivate Meta to explore monetization options for Meta AI in the future.
In summary, Meta’s introduction of a custom search engine into its chatbot ecosystem is a strategic move to enhance functionality while reducing dependence on external services. Through this initiative, Meta aims to better serve its users with more accessible and reliable information.
