Welcome to the July 14, 2025 edition of ACM TechNews, providing timely information for computer professionals three times a week.

HEADLINES AT A GLANCE

• For Algorithms, Memory Is a Far More Powerful Resource Than Time
• ACM President Ioannidis Sees More Humane Role for AI
• Japan Wires the Ocean with an Earthquake-Sensing ‘Nervous System’
• AI Slows Some Experienced Software Developers
• Bodyguards Using Fitness App Revealed Locations of Swedish Leaders
• U.K., France Forge Closer Cyber, Tech Research Ties
• Japan's Smaller Chip Assemblers Join Up to Cut Costs, Share Data
• Voice Assistant Tech for Spacecraft Anomaly Detection, Resolution Tested
• ITU Urges Stronger Measures to Detect AI-Driven Deepfakes
• McDonald's Hiring Bot Exposed Millions of Applicants' Data to Hackers by Using Password '123456'
• AI Challenge to Find Lost Amazonian Civilizations Draws Critics
• Human-like Machine Vision Is More Energy Efficient

For Algorithms, Memory Is a Far More Powerful Resource Than Time

Massachusetts Institute of Technology theoretical computer scientist Ryan Williams recently posted proof of his theory that a small amount of memory is as helpful as a lot of time in all computations. Williams’ proof established a mathematical procedure for transforming any algorithm into a form that uses much less space. Williams' finding has won adulation from the computer science community, including ACM A. M. Turing Award laureate Avi Wigderson, who sent Williams a congratulatory email with the subject line: “You blew my mind."
[ » Read full article *May Require Paid Registration ]

ACM President Ioannidis Sees More Humane Role for AI

The recently-concluded Fourth International Conference on Financing for Development (FFD4) saw ACM host its first Smart Technology, Fair Finance webinar, exploring the benefits and risks of AI in global development finance. Said ACM President Yannis Ioannidis, “Our central message is that technological advancement does not guarantee fair use,” he said. “AI often inspires either fear or unrealistic optimism. In truth, it can do immense good – but only if we ensure it serves everyone, not just the privileged."
[ » Read full article ]

Japan Wires the Ocean with an Earthquake-Sensing ‘Nervous System’

Japan’s new earthquake-detection network, designed to provide early warnings of tsunamis, is comprised of some 3,540 miles of fiber-optic cable zigzagging across 116,000 square miles of ocean that connect 150 observatories on the ocean floor. Each observatory contains 14 sensing channels, including seismometers and accelerometers, as well as pressure gauges to measure waves passing overhead. A tsunami-detection program at Cardiff University in Wales uses machine-learning algorithms to interpret hydrophone signals from ocean stations originally created to listen for clandestine nuclear bomb blasts.
[ » Read full article ]

AI Slows Some Experienced Software Developers

A study by nonprofit METRon on a group of experienced software developers using AI coding assistant Cursor on open-source projects found that the tool actually slowed the developers down when they were working in codebases familiar to them. Before the study, the developers believed using AI would decrease task completion time by about a quarter. The study found, however, that using the AI tool actually increased task completion time by 19%.
[ » Read full article ]

Bodyguards Using Fitness App Revealed Locations of Swedish Leaders

Bodyguards for Sweden’s royal family and Prime Minister Ulf Kristersson uploaded their workout routes to fitness app Strava, an investigation by a Swedish newspaper found, allowing users to map and share their movements. Daily newspaper Dagens Nyheter reported that it tracked more than 1,400 workouts by seven of those bodyguards over the last year, revealing the address of the prime minister’s private residence, the locations of Sweden’s royal family, and the leader of Sweden’s opposition, along with other sensitive information.
[ » Read full article *May Require Paid Registration ]

U.K., France Forge Closer Cyber, Tech Research Ties

A partnership between researchers in the U.K. and France aims to make tech tools and systems used by critical national infrastructure suppliers, including positioning, navigation, and time systems, more resilient against security threats. Additionally, the countries announced a supercomputing partnership between the University of Bristol's Centre for Supercomputing and Grand équipement national de calcul intensif.
[ » Read full article ]

Japan's Smaller Chip Assemblers Join Up to Cut Costs, Share Data

About 30 Japanese companies earlier this year established the Japan OSAT (Outsourced Semiconductor Assembly and Test) Association to bolster the country's back-end semiconductor processing capabilities. These small- and mid-size companies, which handle the final stages of chip production, aim to reduce costs by 20% through collaboration on equipment upgrades and other production procedures. This summer, Japan OSAT plans to launch working groups on automation, data-sharing, and workforce development, among other topics.
[ » Read full article ]

Voice Assistant Tech for Spacecraft Anomaly Detection, Resolution Tested

Texas A&M University researchers are testing the use of virtual assistant technology in space. The Daphne-AT tool uses real-time data streams from a spacecraft's environmental controls and life support systems to determine what is within and outside of baseline tolerances. Explained Texas A&M's Daniel Selva, "Daphne-AT uses a mix of logic and data-driven decision-making to help astronauts make informed decisions when a problem arises on a spacecraft."
[ » Read full article ]

ITU Urges Stronger Measures to Detect AI-Driven Deepfakes

The UN's International Telecommunication Union (ITU) warned in a new report unveiled at its "AI for Good Summit" that advanced detection tools, and social media digital verification and authentication tools, are necessary to address the serious threats to elections and financial systems posed by deepfakes and AI-generated content. Said ITU's Bilel Jamoussi, "Trust in social media has dropped significantly because people don't know what's true and what's fake."
[ » Read full article ]

McDonald's Hiring Bot Exposed Millions of Applicants' Data to Hackers by Using Password '123456'

Independent security researchers Ian Carroll and Sam Curry uncovered major vulnerabilities in the Paradox.ai platform that powers the Olivia chatbot used by McDonald's hiring site, McHire.com. They gained backend access by entering "123456" as a password, revealing a lack of multifactor authentication. Once inside, they accessed a test restaurant's system and applicant data, including logs of every McHire user's chats with Olivia and the personal details of job seekers.
[ » Read full article ]

AI Challenge to Find Lost Amazonian Civilizations Draws Critics

OpenAI's community science project, the OpenAI to Z Challenge, is offering up to $250,000 in cash and credits for OpenAI products to researchers who use the company's tools to search existing Amazon rainforest data and identify lost ancient cities. Archaeologists, Indigenous groups, and other critics say the project violates research norms and ethics codes by failing to consult with the more than 300 Indigenous communities in the Amazon.
[ » Read full article ]

Human-like Machine Vision Is More Energy Efficient

Researchers at Osnabrück University's Institute of Cognitive Science in Germany developed an AI vision model with a structure similar to that of human vision in organization and specialization. Each spatial location in the All-Topographic Neural Network (All-TNN) has its own set of learnable parameters, with a "smoothness constraint" added during training to enable neighboring neurons to learn similar but not identical features. In tests, All-TNN used a tenth of the energy and was three times as strongly correlated with human vision as a convolutional neural network.
[ » Read full article ]