LAION forays into free voice assistance

AI for voice: Large-scale Artificial Intelligence Open Network (LAION), a German non-profit that maintains AI training data sets, has announced a voice assistant which is open to access by everyone.

Called BUD-E, the technology promises an immersive conversational experience better than current AI voice assistants.

So far, Apple’s Siri, Amazon’s Alexa and Google’s voice assistant are the major contenders for AI voice assistance. The technologies have culminated over years since they require extensive infrastructure and R&D investments.

At a time when other projects around voice assistance are mushrooming, Wieland Brendel, a fellow at the Ellis Institute and a contributor to BUD-E told TechCrunch, “Most interactions with [assistants] rely on chat interfaces that are rather cumbersome to interact with, [and] the dialogues with those systems feel stilted and unnatural.”

He said these systems can take commands from controlling music or turning the lights on, but do not have a basis for long and engaging conversations.

LAION’s aim is to create a voice assistant that can manage multi-speaker conversations with interruptions, affirmations and thinking pauses.

AI’s advent is seen in generating codes and responses through text, which is expanding to conversational features introduced by meeting apps like Otter to fish out details and answer conversations in real-time, and like a human, or pitching on behalf of one. 

There have been several issues with generative models for intelligence, ranging from bias to unregulated use spiralling to misuse.

As per a Stanford Study, speech recognition systems from Amazon, IBM, Google, Microsoft and Apple are about twice as likely to mishear Black speakers versus white speakers of the same age and gender, indicating bias.

BUD-E on the other hand is looking at redefining the experience of interacting with voice assistants before generalising that experience to more diverse accents and languages.

The German company has previously been in the news for illegal child abuse images in LAION-5B, having been used by prominent image-generation companies like Stability AI.

It is to be seen how the advent of open technologies like ChatGPT and BUD-E are regulated, even as they promise multifolds of advancement.