Now loading...
Google DeepMind has brought on board the chief executive and a handful of leading engineers from Hume AI, a firm specializing in voice systems that grasp human emotions, through a fresh licensing pact, sources tell WIRED.
While the monetary aspects remain under wraps, Hume AI plans to keep delivering its innovations to competing advanced AI developers.
This arrangement underscores a growing conviction among AI firms that spoken interactions will play a central role in user engagement, with the ability to read vocal tones and sentiments proving essential.
Hume AI anticipates generating 100 million dollars in sales by 2026, collaborating with AI researchers to refine models into more effective audio assistants, according to John Beadle, a cofounder and managing partner at AEGIS Ventures, an early backer of the startup. The company has secured 74 million dollars in investments to date.
Alan Cowen, the CEO with a doctorate in psychology, is moving to Google DeepMind with around seven fellow engineers. They aim to embed audio processing and emotional awareness into the divisions newest systems, insiders explained, requesting anonymity due to speaking without permission.
Hume AI has poured significant resources into crafting algorithms and software for lifelike speech synthesis and identifying feelings from spoken words. Its training involves specialists marking affective signals in actual dialogues. At Google, Cowen and his team will assist in weaving these capabilities into cutting-edge AI frameworks, the sources added.
Voice interaction stands poised to emerge as the main way people connect with AI, with no doubt about that trajectory, stated Andrew Ettinger, a seasoned investor and leader now stepping in as Hume AIs new CEO. He noted that fresh versions of their systems will launch shortly.
Beadle from AEGIS Ventures highlighted that while current AI excels in raw smarts, it lags in truly aiding users by sensing their feelings and adjusting responses to support personal objectives. This gap offers vast potential, especially in everyday gadgets and service interactions.
The move bolsters Googles edge against rivals like OpenAIs ChatGPT, which boasts a natural-sounding speech feature. Google has also inked a long-term collaboration with Apple to infuse its Gemini AI into an upcoming Siri upgrade.
Such pacts like Humes increasingly merge collaboration with talent grabs, letting tech giants secure elite expertise minus the regulatory hurdles of full buyouts. Yet the Federal Trade Commission has signaled plans to examine these so-called aqui-hires more closely.
Last year, Google DeepMind reportedly shelled out 3 billion dollars to access tech from Character.ai, creators of realistic conversational bots. Comparable moves include Microsoft poaching stars from Inflection, Amazon drawing the Adept crew, and Meta luring the head of Scale AI.
