Google Cloud introduces new AI tools for business users
The tech giant released new a translation hub aimed at business users, a computer vision feature for the Vertex AI platform, and the OpenXLA Project open source initiative.
Google Cloud introduced a series of new tools aimed at making it easier for organizations to inject AI into their business ecosystem.
On the first day of its Google Cloud Next '22, the tech giant released Translation Hub, an AI agent that provides users with self-service document translation; new features for its Document AI system; and a new computer vision tool for the Vertex AI automated machine learning platform..
The product releases are designed to help customers "accelerate value capture from AI through the Google Cloud ecosystem," said Chirag Dekate, an analyst at Gartner.
From language translation to image recognition, "a lot of the toolchains that Google is building into Google Cloud ecosystems simplify access [to AI]," he said. "You don't have to think about AI as a standalone component that is implicit within the platform."
Targeting the end user
The release of Translation Hub comes after Google added 24 new languages to Google Translate earlier this year. Expanding on that, Translation Hub lets users translate documents in about 135 different languages within seconds, Google said. Users can also choose a "human-in-the-loop" option that lets local experts edit and review the translated documents for accuracy.
Translation Hub stands out from other translation offerings Google has released in the past, said Dave Schubmehl, an analyst at IDC. Among these are the vendor's translation API aimed at programmers and a mobile translation app.
"Now they've essentially created an end-user-focused [offering] where you don't have to be a programmer or anything like that," Schubmehl said. "It's really meant for the everyday business user."
Translation Hub lets users translate documents in up to 135 languages.
Google also targeted end users for the new features it added to its Document AI system.
"One of the things that people find most difficult right now is how to get all that information out of different types of documents and make it accurate," Schubmehl said.
With the new Document AI Workbench, Google said it's easier for organizations to extract information from a document based on a specific need. Organizations can now extract fields of interest. Document AI Warehouse helps organizations search and manage documents, create workflow controls and manage invoice processing and contracts.
Vertex AI vision and OpenXLA
The Vertex AI Vision service shows the expanding use case for computer vision, Schubmehl said. The computer-vision-as-a-service capability makes computer vision and image recognition AI accessible to data practitioners. Vertex AI vision extends the capabilities of Vertex AI, which was released in 2021 and unified Google Cloud's existing machine learning tools in one environment.
A lot of the toolchains that Google is building into Google Cloud ecosystems simplify access [to AI].
Chirag DekateAnalyst, Gartner
Vertex AI Vision helps developers reduce the time it takes to create computer vision applications by providing a drag-and-drop interface and a library of pre-trained models for tasks like object detection, occupancy counting and product recognition, according to Google.
"These low-cost computer vision APIs have been around for a while, but they weren't always easy to use," Schubmehl said. "Google adding that level of capability is just going to open up more potential use cases for computer vision in the marketplace."
While Google is providing pre-trained models, enterprises that need to custom build their models will have to do a lot more work, he added.
The other challenge is making sure the vision models adhere to responsible AI standards and don't display facial characteristics or private medical records.
Google said it's following its own responsible AI guidance to address some of these issues.
Google also unveiled the OpenXLA Project, an open source machine learning project expected to include participation by other vendors such as AMD, Meta, Nvidia and Intel. It is aimed at helping developers build models with Google's TensorFlow, JAX open source technologies and PyTorch.
"All of these are designed to help enterprises create a pipeline or a project portfolio [so] that they can accelerate production and actually experience benefits of AI," Dekate said.
