Highlights:

  • Google also announced an ambitious project to develop a single AI language model that supports the world’s 1,000 most spoken languages.

At a Google AI event held at the company’s offices at Pier 57 in New York City, Google unveiled several advances in Artificial Intelligence (AI), including generative AI, language translation, health AI, and disaster management.

The program also discussed its attempts to construct responsible AI, especially concerning control and safety, identifying generative AI, and “creating for everyone.”

Google CEO Sundar Pichai said, “We see so much opportunity ahead and are committed to making sure the technology is built in service of helping people, like any transformational technology. Reimagine how technology can be helpful in people’s lives.”

Furthermore, Pichai highlighted the risks and difficulties associated with AI. He stated, “That’s why Google is focused on responsible AI from the beginning, publishing AI principles which prioritize the safety and privacy of people over anything else.”

Google debuts Imagen Video — Phenaki combo

Douglas Eck, principal scientist at Google Research and research director of Google’s Brain Team, discussed a range of Google generative AI announcements, including the company’s cautious, sluggish (relative to DALL-E 2 or Stability AI) attempts to disclose its text-to-image AI systems.

While Google Imagen is not yet available to the general public, the firm stated that it would add a restricted version of the technology to its AI Test Kitchen app (earlier this year demonstrated LaMDA) as a means to collect early feedback. The firm demonstrated a demo, City Dreamer, which allows users to conjure visuals of a city based on a theme, such as pumpkins.

In addition, developing on its text-to-video work unveiled last month, Google released the first rendering of a video that shares both of the company’s complementary text-to-video research methodologies –Imagen Video and Phenaki. The outcome combines Phenaki’s ability to produce a video from a series of text prompts and Imagen’s high-resolution detail.

Douglas Eck, principal scientist at Google Research and research director for Google’s Brain Team, said, “I think it is amazing that we can talk about telling long-form stories like this with super-resolution video, not just from one prompt but a sequence of prompts, with a new way of storytelling.” He further added that he was excited about how filmmakers or video storytellers might use this technology.

Other generative AI advances

In the text area, Eck also presented the LaMDA dialogue engine and the Wordcraft Writers Workshop, which pushed established authors to develop experimental fiction using LaMDA as a tool.

Eck stated that Google would shortly publish a research paper on this topic.

Douglas Eck said, “One clear finding is that using LaMDA to write full stories is a dead end. It’s more useful to use LaMDA to add spice.” He added that the user interface also has to be right, serving as a “text editor with a purpose.”

Eck also highlighted Google’s efforts to use AI to generate code and recently introduced research from AudioLM that extends the audio from any audio clip entered without needing a musical score. It also introduced DreamFusion, the just-announced text-to-3D rendering that combines Imagen and NeRF’s 3D capabilities.

Google is building a universal speech translator

After analyzing Google’s progress in language AI research, Google Brain chief Zoubin Ghahramani outlined the company’s endeavor to represent the diversity of the world’s languages and its ambitious attempt to construct a model that supports the world’s top 1000 languages.

In addition, Google asserts that it is training a universal speech model on more than 400 languages, claiming that this is the “largest language model coverage seen in a speech model today.”

A strong focus on responsible AI

The AI announcements, which included James Manyika, SVP of Google Alphabet, and Marian Croak, VP of engineering at Google, addressed Google’s emphasis on responsible AI.

Croak said, “I think if we’re going to be leaders, it’s imperative that we push state-of-the-art on responsible AI technology. I’m passionate about wanting to discover ways to make things work in practice.”