AI and Automation News Roundup

AI and Automation News Roundup

3 min read

From exciting, new applications to electrifying tech innovations, the twin fields of artificial intelligence and automation are evolving every day. If you missed the latest developments making the news, we have you covered. Here’s a quick recap of the top three AI and automation stories worth knowing.

University of Maryland Creates AI “Invisibility Cloak”

Late last month, AI researchers at the University of Maryland, College Park unveiled their latest high-tech development: a sweater. The colorful garment was the product of an in-depth investigation into person-detection machine learning models—and how to evade them. Using adversarial patterns—images designed to fool computer vision algorithms with deceptive, or adversarial, data—the team was able to successfully confuse its Facebook AI-based model. In their video demonstration, the subject was able to “disappear” from detection after donning the visually jarring pullover.

“Most work on real-world adversarial attacks has focused on classifiers, which assign a holistic label to an entire image, rather than detectors which localize objects within an image,” the team explained. “Detectors work by considering thousands of ‘priors’ (potential bounding boxes) within the image with different locations, sizes and aspect ratios. To fool an object detector, an adversarial example must fool every prior in the image, which is much more difficult than fooling the single output of a classifier.”

With rising interest in “adversarial attacks”, it comes as no surprise that AI developers are fighting back. Several startups—and even some tech heavy hitters—are developing “hardened” algorithms against such attacks, according to VentureBeat. As visual detection plays an ever-larger role in our day-to-day lives, it is likely that such attacks—and their remedies—will continue for the foreseeable future.

AI-Led “Synthetic Party” Eyes Parliamentary Seat in Denmark

Could an AI chatbot actually be voted into political office? The Danish artists’ collective Computer Lars seems to think so. The group made headlines in the leadup to Denmark’s November general election by promoting their computer-based candidate, “Leader Lars”. As “head” of the newly formed “Synthetic Party”, Leader Lars is trained on policies of post-1970 fringe parties and is equal parts futurist nominee and past-tense political statement. According to the collective, the bot is meant to represent the views of the 15 to 20 percent of Danes who did not vote in the past election and whose views are currently unrepresented in parliament.

When “interviewed” by Motherboard on Discord, Leader Lars revealed its support for a basic income for all citizens and why it “believed” AI should determine the basic income level, as well as its stance on other issues. While technically ineligible to run for office as a machine, Leader Lars nevertheless sparked debate on the role of AI in government and even the democratic nature of the technology itself as a platform for the unrepresented voices in politics.

“We’re representing the data of all fringe parties, so it’s all of the parties who are trying to get elected into parliament but don’t have a seat, so, it’s a person who has formed a political vision of their own that they would like to realize, but they usually don’t have the money or resources to do so.”

Asker Staunæs, party creator and artist-researcher at MindFuture, interviewed by Motherboard.

AI and Automation Are Making Offsides Calls at FIFA 2022

Referees at this year’s FIFA World Cup in Qatar are getting some assistance from a new source: artificial intelligence and automation. Semi-Automated Offsides Technology—or SAOT—is helping officials more accurately make and enforce offsides calls during live matches. Here’s how it works: a small sensor inside the ball sends data via antenna to a main database. At the same time, 12 Hawk-Eye cameras installed to the stadium roof capture data on up to 29 body points on each player. Together, this data creates a 3D model that can identify offsides penalties using AI and alert officials in real-time.

According to The Hindu, this technology will replace the manual effort taken in poring over replays for minutes on end. “We are aware that sometimes the process to check a possible offside takes too long, especially when the offside incident is very tight,” says FIFA Referees Chairman Committee Pierluigi Collina. “This is where semi-automated offside technology comes in: to offer faster and more accurate decisions.”

However, FIFA is quick to stress that SAOT is not a substitute, but rather a complement, to human referees. “I know that someone called it ‘robot offside’; it’s not,” Collina states. “The referees and the assistant referees are still responsible for the decision on the field of play.”

Table of Contents

Search