Eternal's new Dark.ai AI assistant is focused on user privacy
What's the story
Eternal has announced Dark.ai, a privacy-focused artificial intelligence platform. The platform will provide encrypted conversations and is expected to be launched globally this year. Dark.ai seeks to tackle data privacy issues that plague mainstream AI assistants, where users often refrain from discussing sensitive topics due to concerns over how their conversations are stored and utilized.
Security measures
Dark.ai's encrypted memory management layer
Dark.ai uses fine-tuned frontier open-weight models on its own infrastructure.
The platform promises end-to-end encryption, ensuring even the company can't access users' conversations.
It also has an encrypted memory management layer that lets context carry across sessions under user control.
This feature is aimed at power users like founders, researchers, scientists, enterprises, and governments who require strict data confidentiality.
Truth over approval
Factual accuracy prioritized over user approval
In addition to privacy, Dark.ai also promises answers that prioritize factual accuracy over user approval.
The models are trained to tell users what they think is the truth rather than simply pleasing them.
Benchmarks comparing their performance with other large language models (LLMs) will be revealed later.
However, details about pricing, commercial model, and technical implementation of its encryption system remain undisclosed.
Market entry
Competition in AI assistant market
The planned launch of Dark.ai comes as tech companies race to build AI assistants for consumers and businesses.
Privacy, reliability, and model performance are key areas of differentiation in this competitive landscape.
For Eternal, Dark.ai marks a foray beyond its core food delivery and quick commerce businesses.
However, the company has not disclosed details about the product's business model or revenue targets yet.