It seems that as we near the end of a banner year for artificial intelligence, the company previously known as Facebook is hard at work, trying to compete in a very critical area – search.
The Information reports that Meta is pursuing its own AI search engine to put inside its Meta AI tool that interacts with users as a powerful large language model.
Here are three things to know about this groundbreaking technology…
It’s Search Inside AI
The first thing to know is that Meta’s goal is to replace search operations inside Meta AI that currently use Google search or Bing.
By creating its own AI search, it will untether its own platform from these traditional sources.
This whole thing makes sense as we start to switch to smart AI tools, and away from traditional tools like Google’s dominant search engine that’s been helping billions of users since the early days of the Internet.
“With new conversational capabilities and image analysis features, Meta AI can take your everyday ideas and make them extraordinary,” write Meta spokespersons as part of the pitch. There’s a lot that a system like this can do to augment human search and more.
Media Deals are Part of the Bargain
Insiders report that as part of this AI search process, Meta has inked a deal with Reuters to have their news appear inside the platform.
Then, too, OpenAI is pursuing its own SearchGPT system, and has apparently partnered with sources like The Atlantic for content.
We can see that companies are wanting users to move into a place where they use AI platforms for search instead of traditional browser search engines.
Assistants and Agents
This phenomenon is also just one small part of the ‘assistant to agent’ model happening right now.
We’ve had a lot of work on AI assistant technologies that are more vibrant than traditional offerings, but not autonomous in the way that the newest systems are.
We reported, for example, on how liquid AI models are driving more intelligence to the edge, and how they may end up endowing new models with forms of reasoning.
We also pointed to the emergence of OpenAI’s o1 model, which has its own reasoning capabilities like error correction, and an even newer Orion model that’s supposed to be debuted this year or next year.
And that’s not the only company working on its next-generation models – there’s Anthropic’s Claude and Google Gemini, and a handful of other efforts that are moving the ball forward.
In general, experts are talking about the power of ‘agentic AI’ – that combines its formidable knowledge base with proactive initiatives. For example, Claude can now use a personal computer in the same way that a human can. That means instead of looking over the agent’s shoulder, the human can give a lot more responsibility to the AI, and sit back and let it do its work.
“Agentic AI refers to a new class of artificial intelligence systems designed to act with autonomy, making decisions and taking actions without direct human intervention,” writes Kate Rogerson this month at Interface.ai. “These systems are capable of processing vast amounts of data, reasoning, and adapting to real-time changes in their environment. The evolution from traditional AI to agentic AI signifies a leap forward in how systems can handle complex tasks, such as reasoning, planning, and learning from experience – much like a human agent.”
So look for these kinds of models to start popping up in use cases in various industries. Our world will never be the same.
