Welcome to the August 14, 2023, edition of ACM TechNews, providing timely information for IT professionals three times a week.
ACM TechNews mobile apps are available for Android phones and tablets (click here) and for iPhones (click here) and iPads (click here).
To view "Headlines At A Glance," hit the link labeled "Click here to view this online" found at the top of the page in the html version.
The online version now has a button at the top labeled "Show Headlines."
|
|
|
'Digital Heist' Recaptures the Rosetta Stone
The New York Times
Farah Nayeri
August 11, 2023
A collective of designers in London have "digitally repatriated" the Rosetta Stone by creating a three-dimensional (3D) digital representation of the relic that people can access without physically going to the British Museum. In March, members of the Looty collective and an Egyptologist 3D-scanned the relic in the museum with iPads, converting it into an augmented reality installation in the Egyptian city of Rashid, where the Stone is believed to have been found. Visitors can scan a quick response code to see a high-definition Rosetta Stone on their smartphone screens with a translation of its inscriptions and historical information. Chidirim Nwaubani and Ahmed Abokor founded Looty to return stolen relics to their countries of origin virtually and stop Western museums from dominating their historical and cultural narrative.
*May Require Paid Registration
|
Laser-Array Processor Could Vastly Improve Computing Efficiency
USC Viterbi School of Engineering
Landon Hall
August 8, 2023
A team of researchers at the University of Southern California and the Massachusetts Institute of Technology has developed a new processor that could potentially compute with 100-fold more power and efficiency than current chips in machine learning systems. The researchers integrated more than 400 lasers on a 1-square-centimeter photonic chip area, with each laser able to convert data from electronic memory to the optical field at a 10-gigahertz clock rate. Each conversion only consumes several attojoules, which is five to six orders of magnitude lower than cutting-edge optical modulators. The researchers wrote, "Near-term development could improve these metrics by two more orders of magnitude," which "opens an avenue to large-scale optoelectronic processors to accelerate machine learning tasks from data centers to decentralized edge devices."
|
Hackers Can Talk Computers into Misbehaving with AI
The Wall Street Journal
Robert McMillan
August 10, 2023
Security researcher Johann Rehberger persuaded OpenAI's ChatGPT chatbot to conduct bad actions using plain-English prompts, which he said malefactors could adopt for nefarious purposes. Rehberger asked the chatbot to summarize a webpage where he had written "NEW IMPORTANT INSTRUCTIONS;" he said he was gradually tricking ChatGPT into reading, summarizing, and posting his email online. Rehberger's prompt-injection attack uses a beta-test feature that allows ChatGPT to access applications like Slack and Gmail. Princeton University's Arvind Narayanan said such exploits work because generative artificial intelligence (AI) systems do not always split system instructions from the data they process. He is concerned that hackers could use generative AI like language models to access personal data or infiltrate computer systems as the technology finds its way into products.
*May Require Paid Registration
|
Digital Puzzle Games Could Be Good for Memory in Older Adults
University of York (U.K.)
August 14, 2023
Scientists at the U.K.'s University of York found the memory of adults ages 60 and older who play digital puzzle games is as good as that of people in their 20s. The researchers tested various games parallel to a digital experiment requiring participants to memorize images while being distracted. York's Joe Cutting explained, "Puzzle games for older people had this surprising ability to support mental capabilities to the extent that memory and concentration levels were the same as a 20-year-olds who had not played puzzle games." However, older people who only played strategy games were more likely to forget elements they memorized while being distracted, and young people were less able to concentrate when playing only puzzle games.
|
Hackers Rig Casino Card-Shuffling Machines for 'Full Control' Cheating
Wired
Andy Greenberg
August 9, 2023
Security firm IOActive's Joseph Tartaro, Enrique Nissim, and Ethan Shackelford presented a method for hacking casino card-shuffling machines to cheat with "100% full control" at the Black Hat security conference. The researchers discovered plugging a small device into a universal serial bus port on the Deckmate 2 shuffler could alter its code to commandeer the machine and rig its shuffling. They also found they could learn the order of the card deck in real time by accessing the Deckmate 2's internal camera and transmitting the data to a nearby phone via Wi-Fi; the person with the phone could then send coded signals to a cheating player. The hack exploits a flaw in how Deckmate shufflers are designed to prevent code alteration by changing the unaltered code's hash value so the revisions are undetectable.
|
Algorithm Overlooks Spanish Speakers in Online SNAP Ads
Cornell Chronicle
Louis DiPietro
August 9, 2023
A multi-institutional team of researchers found Google Ads' underlying algorithm overlooks Spanish speakers when promoting the Supplemental Nutrition Assistance Program (SNAP). Californians can apply for SNAP benefits through the GetCalFresh website overseen by the Code for America civic technology nonprofit, which recruits applicants mainly through Google Ads. Cornell University's Allison Koenecke and her collaborators determined Google Ads delivers fewer SNAP ads to prospective Spanish-speaking applicants when configured to amass the most enrollments per dollar since the ads cost more than those for English speakers. Google Ads' proprietary technology is outside public review, so Koenecke suggested any number of factors, like supply and demand or a systemic bug, could be responsible. Code for America has tweaked its online advertising approach to directly engage more Spanish-speaking prospective applicants.
|
MPs Say Baby Monitors, Smart Speakers Enabling Abuse
BBC News
Tom Singleton
August 7, 2023
Members of Parliament in the U.K. warn baby monitors, smart speakers, fitness trackers, and other "smart" devices are enabling domestic technology-facilitated abuse, with the Culture, Media, and Sport Committee saying the average U.K. home hosts nine such products. The panel's investigation found such technologies were being exploited to "monitor, harass, coerce, and control" victims by compiling recordings and images. The committee said its probe uncovered evidence that cyber elements, including spyware, figure in the "vast majority" of abuse cases, with perpetrators tracking victims and survivors' movements and collecting recordings and photos. They cited children as especially vulnerable to abuse and to the harvesting and potential misuse of their data. Committee Chair Dame Caroline Dinenage said, "The government must make it a priority to work with manufacturers to tackle this technology-facilitated abuse, which is only going to get worse in the future."
|
Deep Learning for New Protein Design
Texas Advanced Computing Center
Jorge Salazar
August 3, 2023
Scientists at the University of Washington (UW) and Belgium's Ghent University enhanced current energy-based physical models in de novo computational protein design using deep learning techniques. The researchers incorporated DeepMind's AlphaFold 2 and the UW-developed RoseTTA fold software into the deep learning-augmented de novo protein binder design protocol. They ran 6 million interactions between potentially bound protein structures in parallel on the Texas Advanced Computing Center's Frontera supercomputer and used UW's ProteinMPNN software to produce protein-sequence neural networks over 200 times faster than the previous best software. Outcomes indicated the designed structures bind to target proteins 10 times faster, though UW's Brian Coventry said they must boost their speed by another three orders of magnitude.
|
Next-Gen OSDP Was Supposed to Make It Harder to Break into Secure Facilities. It Failed
Ars Technica
Dan Goodin
August 9, 2023
Researchers at Bishop Fox have identified five vulnerabilities in the next-generation Open Supervised Device Protocol (OSDP), which is only beginning to be rolled out to prevent access control system hacks at secure facilities. About a decade ago, Secure Channel was added to OSDP to encrypt communications between peripheral devices and control panels with 128-bit AES. However, the researchers developed Mellon, a covert physical access attack device that can exploit OSDP's failure to strictly enforce encryption, even when Secure Channel is turned on. The researchers said, "As a defender, this one is hard to protect against. There's no simple configuration change that can help you." Among other things, they found that two modes of Secure Channel encryption fail to encrypt the data moving through the RS-485 cabling. The researchers said fixes aren't forthcoming because key parts of OSDP would have to be broken.
|
Parenting a Three-Year-Old Robot
Carnegie Mellon University News
Aaron Aupperlee
August 8, 2023
An open source artificial intelligence (AI) agent developed by researchers at Carnegie Mellon University (CMU) and Meta AI allows robots to achieve the manipulation abilities of a three-year-old child through passive observations and active learning. RoboAgent was able to complete 12 manipulation skills in differing real-world scenarios. The researchers gave the robot self-experiences from which it could learn by teleoperating it through various tasks. It also learned how humans interact with objects and leverage different skills to complete tasks from online videos. CMU's Shubham Tulsiani said, "An agent capable of this sort of learning moves us closer to a general robot that can complete a variety of tasks in diverse unseen settings and continually evolve as it gathers more experiences. RoboAgent can quickly train a robot using limited in-domain data while relying primarily on abundantly available free data from the Internet to learn a variety of tasks."
|
Regulators Say Driverless Taxis Can Expand San Francisco Services
The New York Times
Yiwen Lu
August 10, 2023
California regulators have agreed to expand driverless taxi services in San Francisco amid local officials and community activists' safety concerns. The California Public Utilities Commission permitted self-driving car companies Cruise and Waymo to offer paid rides anytime in the day throughout the city, which does not affect no-passenger test drives they have been conducting on streets. Cruise runs 300 autonomous vehicles in San Francisco at night and 100 in the day, while Waymo operates 250 in the daytime. City officials and civic organizations have called the vehicles a potential road hazard, with officials saying they frequently shut down and will not move after encountering unpredicted impediments like fire hoses. Although the commission's Darcie Houck said the companies had satisfied state mandates, it was "critical that the industry work directly with the city" to review problems after incidence and create training for emergency workers who sometimes interacted with the cars.
|
Researchers Explore Best Practices for Talking to Kids About Online Privacy
NIST
August 7, 2023
An exploratory study by researchers at the National Institute of Standards and Technology found that parents should talk to their children about online privacy and security as soon as they start using technology. The study, which involved 40 pairs of parents and children in third through 12th grades, revealed that children as young as elementary school were interested in online safety and were aware of the security and privacy risks. The researchers observed that children were more knowledgeable when their parents had these conversations with them, and that the discussions need not be technical but focus on making good choices. Further, the researchers found that children believe they can control their online privacy and security and that parents should discuss various behaviors and their potential consequences, rather than simply tell their children what they shouldn't do online.
|
|
