This week, Google reveals new devices, including the new Pixel 3 smartphones and Pixel Slate tablet and news about an Amazon experiment using artificial intelligence to screen job applicants. But first, Google+ gets shuttered.

Google shuts down Google+ after Wall Street Journal uncovers vulnerability

An article in the Wall Street Journal last week asserted that the private information of users of Google+ had been vulnerable without the company informing them.

Google has since decided to close the Google+ social networking service.

The report alleges that between 2015 and March of this year, a software vulnerability in the Google+ code meant that user profile data was accessible by developers outside Google.

There is no evidence, Google says, “that any developer was aware of this bug, or abusing the API, and we found no evidence that any Profile data was misused.”

The software bug was fixed in March 2018. Google’s bigger problem is that the company kept it all a secret.

Internal memos discussing the problem and the way to handle it were source documents used in the reporting by the Wall Street Journal. “The document shows Google officials knew that disclosure could have serious ramifications. Revealing the incident would likely result ‘in us coming into the spotlight alongside or even instead of Facebook despite having stayed under the radar throughout the Cambridge Analytica scandal,’ the memo said. It ‘almost guarantees Sundar will testify before Congress.’”

This comes on the heels of news in September that 50 million Facebook users may have had their accounts accessed by hackers.

Pixel 3 smartphone, Pixel Slate, and other products announced by Google

At an event in New York yesterday, Google showed off its new smartphones, the Pixel 3 and Pixel 3 XL.

Priced starting at $999 for the 5.5-inch Pixel 3 and $1129 for the 6.3-inch Pixel 3 XL with 64 GB of storage, Google is playing up the impact of artificial intelligence with these new devices. The company says that AI is being used to prioritize battery use to help get you through the day, to suggest phrases so writing emails is quicer, and to improve the photos you take.

Speaking of photos, the Pixel 3 handsets have two front-facing cameras, one with a wide-angle lens, to give you more selfie flexibility (“Groupie” anyone?).

The single camera on the back has a resolution of 12.2-megapixels. And Google is offering free cloud-based storage in your Google Photos account for all photos and videos that you take with your Pixel 3, at their original resolution (including 4K).

The Pixel Stand ($109), meanwhile, is a wireless charging stand. It was also announced today, and while it’s designed by Google ostensibly for using with the Pixel handset, it uses the Qi standard and so can be used with devices from any manufacturer. It supports fast charging and is powered by a USB-C cable.

Pre-order either of the two Pixel 3 models – they release on October 18 – and Google will throw in a Pixel Stand.

The Pixel Slate is Google’s latest take on a tablet-keyboard combo. The Pixelbook, released last winter, is a Chromebook with a built-in keyboard equipped with a 360-degree hinge, but the Slate is more of a tablet, and has a cheaper price point as a result.

Starting at $849 and releasing later this year, the Slate runs the Chrome OS operating system, which has been redesigned with touch screens in mind.

The Slate Keyboard ($259) and Pixelbook Pen ($129) are extra.

Amazon’s failed experiment using AI to screen job applicants

As reported by Reuters today, Amazon at one point was using machine learning to rate job candidates. The problem was that the information used to train the machines were resumes previously submitted, which tended to come from male applicants.

So the AI learned that males were preferred, and was “penalizing” and “downgrading” female applicants.

The company is no longer using the system and sources said the AI recommendations were not paramount when making hiring decisions.

The experiment serves to highlight the big constraint on AI and machine learning: They are only as good as the information they learn from.

Recall the MIT researcher who discovered that facial recognition software in use around the world is racially biased.