Thumbnail not available.
Click to view the full image.

Here’s what you need to know about Rai, Rappler’s conversational chatbot.

More details about Rai and why Rappler developed Rai on this story https://www.rappler.com/about/rai-artificial-intelligence-conversational-chatbot-journalist-designed/

Yes, Rai gets the latest articles every 15 minutes from the Rappler website!

A lot of guardrails, which in layman’s terms are telling Rai to behave a certain way, and by constricting its data sources to trusted and curated facts.

Way before we developed Rai, Rappler started developing its own Knowledge Graph, which Rai is able to access using the GraphRAG method. 

This helps Rai respond to questions more accurately.

GraphRAG means:

Graph = Knowledge Graph

RAG = Retrieval Augmented Generation

As in all content published by Rappler, Rai's responses are governed by our Corrections policy. If you found an error in Rai's responses, kindly email feedback@rappler.com. We will endeavor to root out the cause of the mistake and correct the source material if needed.

For more on Rappler's editorial standards and guidelines, including our corrections policy, please refer to the link below:

https://www.rappler.com/about/standards-guidelines-corrections-fact-check-content-comment-moderation/

Did you know that Rappler is the first newsroom in the Philippines to have published its AI guidelines?

You can access the guidelines through the link below:

https://www.rappler.com/about/guidelines-artificial-intelligence-usage-development/

A knowledge graph is a kind of database that is able to capture relationships between various real-world entities: for instance, people, places, events and other news topics. This makes it possible to visualize all of the information in a network. 

Most databases that people use are more like a row structure, this is a series of graph and nodes, like a tree.

Not at all. However your questions are on the chat room and the AI research team uses that to improve Rai currently.

Nope! We only pass the chunks needed so that Rai can formulate the answer, protecting the integrity of our data.

You can ask Rai questions about people, places, events and concepts that figured in Philippine news and more. Rai also has access to news and information from other countries but these are typically presented from the perspective of the Philippines.

Not in the least! 

Conversational Rai is vastly different from Rai, the FGD moderator that Rappler introduced at the 2023 Social Good Summit. 

Conversational Rai involves a complex architecture which can be connected with any of the current large language models. This new architecture includes data pipelines and swappable AI models (So yes, we can use OpenAI’s models, but our approach is model agnostic.) 

But while we use the language generation powers of existing models, Rai’s source of truth is Rappler’s stories and a knowledgebase sourced from databases gathered, processed, and vetted by Rappler’s researchers.

Retrieval augmented generation or RAG is a method for enhancing the accuracy and reliability of an AI model by grounding it using an external source of knowledge or data.

Rai is incredibly powerful and has a multitude of possible use cases:

💠 Fighting disinformation by always providing sources for its answer

💠 Sifting through difficult sources and complex datasets, helping journalists (and the public) find the root of the story

💠 Civic engagement and promoting civic discussion

💠 Ending obfuscation in bureaucratic processes, where people try to “hide” anomalies in paperwork - Rai can help you understand the data!

💠 Understanding your world

💠 Teaching students difficult topics

💠 Talking in different languages and helping people find a common ground - ask Rai a question in any supported language to see how it works! Cebuano, Japanese, it can do it all!

💠 Show you how people think and creating a place where it feels safe to ask questions in the open

💠 Understanding the news in a flash

Rai will evolve into a much more amazing form soon, where it can make connections that users wouldn’t normally make.