Google announced an extensive array of new AI models during its I/O developer showcase on Tuesday.
With about 13 distinct AI tools unveiled during the event, the tech giant demonstrated the extent of its AI-focused research.
Some of the more interesting product updates/launches include:
- Project Astra, a project that uses a phone's camera and AI to locate objects
- Circle to Search, a gesture-based search feature
- TalkBack, a screen reader to assist low-vison users
Google also introduced Veo, a tool that is capable of generating 1080p resolution videos in different visual and cinematic styles, rivaling OpenAI’s text-to-video model Sora.
In a promotional video, actor-comedian-rapper Donald Glover and his creative studio Gilga used Veo to create a short film.
"Everybody's gonna become a director, and everybody should be a director. Because at the heart of all of this, is just storytelling," the "Mr. and Mrs. Smith" star says.
"The closer we are to being able to tell each other our stories, the more we'll understand each other," he adds.
Veo's selling point is that it is better equipped to understand and execute nuanced prompts in a more cinematic way.
Fighting AI Abuse
One of the biggest concerns in the release of new generative AI models is how they can be abused to create misinformation and nonconsensual sexual content like deepfakes.
This was addressed when Google’s DeepMind CEO Demis Hassabis introduced SynthID during his keynote presentation.
Originally launched in August, SynthID is now equipped with better technology that allows watermarks to be embedded in AI-generated images, which are identifiable by the system even though they are undetectable to human sight.
The model has also been adapted to insert inaudible watermarks into AI-generated music made with DeepMind’s Lyria model.
Hassabis also demonstrated how SynthID could now watermark videos created by Veo and texts generated by Google's language models, allowing for deepfakes to be spotted more easily.
Advancements in AI safeguards like watermarking are part of a broader effort to uphold ethical standards, with the Biden administration urging federal agencies to create guidelines to address AI misuse.
READ NEXT: OpenAI's GPT-4o Allows You to Converse with AI
Editing by Katherine 'Makkie' Maclang
