Addressing “the worry and the opportunity”


25 July 2018
Google Next 2018

“Security is the no.1  worry and AI is the no. 1 opportunity for many companies,” said Diane Greene, chief executive officer of Google Cloud. And the company is addressing these with a slew of product announcements at the Google Next 2018 conference in San Francisco.

Addressing the security worry

On the security front, Google announced new moves to secure access to apps and services with the introduction of context-aware access and the Titan Security Key as a second factor of authentication.

Context-aware access allows organisations to define and enforce granular access to Google Cloud Platform (GCP) application programming interfaces (APIs), resources, G Suite, and third-party Software as a Service (SaaS) apps based on a user’s identity, location, and the context of their request. The tech giant says this increases an organisation’s security posture while decreasing complexity for its users, giving them the ability to seamlessly log on to apps  from anywhere and any device.

Context-aware access capabilities are available for select customers using ​VPC Service Controls​, and are coming soon for customers using ​Cloud  Identity and Access Management​ (IAM), ​Cloud Identity-Aware Proxy​ (IAP), and ​Cloud Identity​.    

Another security announcement was that of the ​Titan Security Key​, a ​FIDO​ security key that includes firmware  developed by Google to verify its integrity. “We’ve long advocated the use of security keys as  the strongest, most phishing-resistant authentication factor for high-value users, especially  cloud admins, to protect against the potentially damaging consequences of credential theft,” said Jennifer Lin, Product Management Director, Google Cloud, in a blog post on the security announcements.    

Titan Security Keys are available now to Google Cloud customers and will soon be available for purchase on the Google Store.   

Other security launches at Google Next 2018 included:

  • Shielded Virtual Machines (beta), which helps to ensure that VMs have not been tampered with, by enabling orgnaisations to monitor and react to any changes in the VM baseline as well as its current runtime state.
  • Binary Authorisation (coming soon to beta) which enables organisations to  enforce signature validation when deploying container images.
  • Container Registry Vulnerability Scanning (coming soon to beta) which automatically performs vulnerability scanning for Ubuntu, Debian and  Alpine images to ensure that images are safe to deploy.
  • Cloud Armor geo-based access control (beta), a distributed denial of service (DDoS) and application defence service which allows organisations to control access to their services based on the geographic location of the client trying to connect to their application.

Democratising AI

At Google Next 2018, the company also announced new products aimed at “democratising AI”.

Fei-Fei Li, Chief Scientist of AI and Machine Learning, said AI was no longer a niche for the tech world but a differentiator for businesses in every industry.

Earlier this year, Google launched AutoML, a suite of machine learning products aimed at extending ML models to recognise data specific to different organisations without them having to write code. “With AutoML, anyone can make machine learning work for them,” she said.

The first release, AutoML Vision, provides cloud vision APIs for businesses to train image recognition models. In the latest iterations, AutoML has been extended to cover natural language and translation capabilities.

AutoML Natural Language builds on Google’s text natural language API to understand subject matters specific to particular industries. AutoML translation extends ML translation technology to recognise jargon, terms and figures of speech in 27 language pairs.

Othe AI/ML related announcements at the conference included

  • New features in Dialogflow Enterprise Edition to help organisations build AI-powered virtual agents for the contact centre, including phone-based conversational agents (interactive voice response or IVR).
  • Contact Centre AI, which allows enterprises with limited ML learning expertise to deploy AI in their contact centres.
  • BigQuery ML (beta), which enables data analysts and data scientists to build and deploy ML models using SQL statements.
  • Extension of Cloud Vision API to recognises handwriting, additional file types (PDF and TIFF) and product search, and identify where an object is located within an image.
  • Extension of Cloud Text-to-Speech to provide multilingual access to voices generated by DeepMind WaveNet technology and optimise responses for the type of speaker, and
  • Extension of Speech-to-Text to identify what language is spoken as well as different speakers in a conversation.