Google has taken the wraps off a range of AI services it says will allow companies to develop and deploy machine learning and AI models – depending on their comfort level with using beta services.
AI Platform, released in beta, is described as a “comprehensive, end-to-end development platform that helps teams prepare, build, run and manage ML products via the same shared interface.”
The search giant says it will allow teams to collaborate on model sharing, training and scaling workloads from a shared dashboard in the Cloud Console. The platform has a built in labelling service for training data, including images, video, audio and text. It supports Kubeflow, which Google says means ML pipelines are portable, and can be run either on premises or in the cloud with minimal changes.
Google’s Cloud AutoML service for organisations with minimal machine learning expertise also gets a stack of new features.
The first new addition is AutoML Tables, which arrives in beta, and promises the ability to create machine learning models from tabular datasets “with zero code” – if the data is in BigQuery or Google’s cloud storage services.
Also in beta is AutoML Video, which promises automatic classification of video using user defined labels. The service promises to be easier to use than Google’s Video Intelligence API, albeit without many of the tracking and analysis features the latter promises. In a similar vein, the AutoML Vision Edge service targets the creation of models on connected edge devices.
Google has launched its Document Understand AI service in beta, which it describes as “a scalable, serverless platform to automatically classify, extract, and enrich data within your scanned or digital documents”. Google said that it will integrate with technology stacks from “partners and third parties” and mentioned it is already being used by organisations like Iron Mountain and Box – meaning your most ancient documents could be serving up insights. Or not.
Getting even more business specific, Google has launched a beta of the Contact Centre AI service it introduced last year, which includes updates to voice models. It said Avaya, Salesforce and Accenture were partnering on the service.
The search giant has also trained its eye on the retail sector, with its Vision Product Search now generally available for retailers to build into their mobile apps. It has also launched a beta of Recommendations AI, designed to help retailers make recommendations to customers to “drive engagement and growth”, ie, sell stuff.
While a certain Seattle-based company might have a much larger share of the market for AI-based cloud services, Google does have one advantage when it comes to retail – its customers won’t see it as an existential threat.
Conversely, it will be interesting to see how the vendor putting AI and ML tech in the hands of customers plays with its own workers. It has already been forced to back off one deal with the US Defence Department due to worker pressure, while an attempt to put together an AI advisory panel foundered just last week. Will Google’s workers object to third party companies taking Google’s tools to go and construct their own ethically questionable AI or ML projects? That is difficult to predict.