Sendbird Becomes the First Communications API Platform to Integrate Open-Source LLM to Address AI Chatbot Enterprise Privacy Concerns

Llama 2 integration makes it possible for Sendbird to further empower companies to leverage generative AI securely without exposing data to third-party servers.

Sendbird, the industry-leading communications API platform for web and mobile apps, today announced a groundbreaking integration to revolutionize how businesses handle generative AI-powered customer conversations. Sendbird has integrated Llama 2, Meta’s Open-Source Large Language Model (LLM), into its AI-powered communications platform.

By adding support for Llama 2, on top of existing ChatGPT and PaLM-2 integrations, Sendbird becomes the first company to enable customers to leverage the power of generative AI chatbots in a manner that best suits their overall strategy and security preferences. With Llama 2 as an open source option, Sendbird allows customers unprecedented transparency, control, and flexibility over their communications experiences.

“Sendbird is known for its dedication to enhancing communication in web and mobile apps and for harnessing the power of generative AI to benefit business. Now, we are taking another critical privacy-oriented step by integrating an open-source LLM on our platform,” said Sendbird CEO and Co-founder John S. Kim. “By hosting the Llama 2 LLM within the Sendbird infrastructure, enterprises can now retain their conversations’ data on the Sendbird servers, ensuring the utmost security and reducing the risk of unauthorized access or data leaks.”

Marketing Technology News: Bloomreach Announces Bloomreach Search and Discovery Cartridge for B2C Commerce on Salesforce AppExchange, the World’s Leading Enterprise Cloud Marketplace

The Importance of Open-Source

Open-Source LLMs, epitomized by Meta’s Llama 2, represent a monumental leap forward. These AI language models, trained by billions of words, are free and accessible for companies to adapt them to their needs.

Additionally, open-source LLMs provide a new level of transparency. This is imperative when LLMs are known to pose risks, as they can memorize sensitive data from training sets, which malicious actors can then exploit. With open-source LLMs, like Llama 2, companies have clear insight into the LLM architecture, training methodologies, and usage. Based on this information they can decide how such models are used. It also helps build trust, facilitate audits, and ensure ethical and legal compliance.

Sendbird is developing fine-tuning expertise and will provide support capabilities to enterprise companies looking to enhance the LLM performance with their proprietary data while maintaining data privacy, governance, and security.

“We look forward to harnessing the benefits of open-source LLMs for our chatbots. This addition not only brings customers new levels of control and security, it provides choice and flexibility for implementing generative AI,” added Kim. “From large-scale gaming and social and community businesses to food delivery, healthcare, financial service organizations, and enterprises worldwide, Sendbird powers meaningful communications and customer interactions at scale. We will continue to innovate across our platform to ensure our customers always have the most advanced and useful tools at their disposal.”

Marketing Technology News: MarTech Interview with Tobyn Sowden, CEO at Redbrick