Latest Instagram Update: AI for Bullying
At the Marketing AI Institute, we read dozens of articles on artificial intelligence every week to uncover the most valuable ones for our subscribers and we curate them for you here. We call it 3 Links in 3 Minutes. Enjoy!
Instagram Using AI to Proactively Detect Bullying
The Verge covered Instagram’s first product announcement made under new chief Adam Mosseri: AI for cyberbullying.
Last year, Instagram began leveraging machine learning to watch for offensive language used in photo captions. Now, this updated AI can “view” images to check for bullying, as well.
Instagram hasn’t yet shared details on exactly what the AI will look for in images. However, one example given so far is the machine’s ability to look for split-screen images, as these typically compare photos and can be negative.
While AI has the capability to check through millions of images faster than humans, it is only the first line of defense against bullying. Since AI tools don’t “read” the same way humans do, they often have trouble understanding human context and nuances. Once the program has detected a potentially offensive or harassing image, it is sent to a team of human moderators for final review.
This update will be live in time for the US’s National Bullying Prevention Month in October and Anti-Bullying Week in the UK.
Amazon Scraps Biased AI Recruiting Tool
This week Amazon announced it is scrapping a secret AI recruiting tool due to its bias against women.
According to CNBC, Amazon’s machine learning specialists have been building computer programs to aid in hiring since 2014. The goal was to liken the recruiting process to product reviews—the machine would digest resumes and give each a rating from one to five stars based on job descriptions.
Unfortunately, the computer models weren’t rating candidates for software developer jobs and other technical posts in a gender-neutral way.
The tool had been trained to observe patterns in resumes submitted to Amazon in the past 10 years. Because the tech industry has been a primary male-dominated field, most of these resumes came from men. As a result, the AI downgraded resumes with the word “women’s” (as in “women’s chess club captain”) and graduates of all-women colleges.
Amazon has attempted to edit the programs to remove gender biases. However, with no guarantee that the machines are completely void of discrimination, the project team has disbanded.
Google Pixel 3 Loaded With AI Features
Google announced a full lineup of new products at its annual hardware launch yesterday. The Pixel 3 smartphones stole the show with their complete build-out of AI capabilities. Our favorite highlights from Inverse are below.
Pixel 3 phones can intercept spam calls with an AI assistant that answers suspicious calls and transcribes the conversations, so you never have to pick up the phone.
Another AI-powered feature is the integrated Google Lens, which enables users to point their Pixel 3 camera at items (shoes, albums, etc.) and get search results for similar products immediately—all without typing a single word.
Debatably our favorite feature is Google Duplex, Google’s next-generation voice assistant. This upgraded AI can book appointments, make reservations and complete your most repetitive tasks on your behalf.
Google’s senior vice president of hardware, Rick Osterloh, emphasized the company’s belief in artificial intelligence: “Our deep investment in A.I. is at the core of our ability to push forward in many areas at once and you’re going to see this in a lot of our products today. When we think of artificial intelligence in a consumer hardware context, it isn’t artificial at all. It’s helping you get real things done every day.”