Welcome to the April 9, 2025 edition of ACM TechNews, providing timely information for computer professionals three times a week.

HEADLINES AT A GLANCE

• Cong Receives ACM Breakthrough in Computing Award
• White House Orders Agencies to Develop AI Strategies
• DOGE Is Coming for U.S. Government’s Magnetic Tapes
• AI Could Impact 40% of Jobs Worldwide in the Next Decade, U.N. Agency Warns
• Meta Draws Congress’ Scrutiny over China Claims
• Researchers Uncover Hidden Patterns Behind $3.5-Billion Cryptocurrency Collapse
• U.S. AI Lead over China Rapidly Shrinking
• Apple's Encryption Row with U.K. Should Not Be Secret, Court Rules
• Taiwan Says China Uses Generative AI to Ramp Up Disinformation
• NIST Calls Time on Older Vulnerabilities Amid Surging Disclosures
• Bringing Personal Items to Mixed Reality
• Bacteria-Inspired Robot Roams Underwater
• Researchers Investigate AI Threats in Software Development

Cong Receives ACM Breakthrough in Computing Award

ACM Fellow Jingsheng Jason Cong was named to receive the 2024 ACM Charles P. “Chuck” Thacker Breakthrough in Computing Award for fundamental contributions to the design and automation of field-programmable gate arrays (FPGAs) and customizable computing. The ability to change the functionality of FPGAs after manufacturing has made them part of standard hardware in datacenters, telecommunications, aerospace, defense, and automotive engineering. Said ACM President Yannis Ioannidis, “Cong’s work in enabling integrated circuits to be programmed, or re-programmed, led to a new level of versatility, efficiency, and power for a wide range of applications."
[ » Read full article ]

White House Orders Agencies to Develop AI Strategies

An April 7 memo from the White House Office of Management and Budget gave government agencies six months to "develop an AI strategy for identifying and removing barriers to their responsible use of AI and for achieving enterprise-wide improvements in the maturity of their applications." The memo also instructed federal agencies to name chief AI officers and establish generative AI policies. A separate White House directive called for "efficient acquisition of [AI] in government" and for agencies to focus on interoperability and "maximize the use of American-made AI."
[ » Read full article ]

DOGE Is Coming for U.S. Government’s Magnetic Tapes

The U.S. Department of Government Efficiency (DOGE) this week tweeted it had saved $1 million by converting 14,000 records on tape at the U.S. General Services Administration to “permanent modern digital records.” Tape, they noted, was “70-year-old” technology. The tweet has sparked debate among government technologists, some of whom noted that if stored properly, tapes last longer than hard-disk drives. They are also secure from hacking, and require no power to maintain.
[ » Read full article *May Require Paid Registration ]

AI Could Impact 40% of Jobs Worldwide in the Next Decade, U.N. Agency Warns

According to a report from the U.N. Department of Trade and Development, 40% of jobs across the globe could be impacted by AI in the next decade. The report also noted that nearly half of global research and development spending in AI can be attributed to 100 companies, most based in the U.S. and China.
[ » Read full article ]

Meta Draws Congress’ Scrutiny over China Claims

U.S. lawmakers are investigating a claim by a former Facebook employee that Meta CEO Mark Zuckerberg lied to Congress about the company’s efforts to launch its social network in China. The probe is focused on alleged work to censor content and provide AI tools, including surveillance software, to the Chinese Communist Party. The censorship efforts “allegedly extended to dissidents outside of China, including in the U.S.,” the lawmakers wrote, citing internal documents.
[ » Read full article ]

Researchers Uncover Hidden Patterns Behind $3.5-Billion Cryptocurrency Collapse

Researchers from the U.K.'s Queen Mary University of London and Pometry, a university spinoff company, developed software that identified patterns behind the May 2022 collapse of the TerraUSD stablecoin and its associated currency, LUNA. The researchers used temporal multilayer graph analysis to map the relationships between different cryptocurrencies traded on the Ethereum blockchain. They found five or six traders accounting for nearly all the trading activity on key days, which they said indicated the traders likely colluded to trigger the collapse.
[ » Read full article ]

U.S. AI Lead over China Rapidly Shrinking

Stanford University's latest Artificial Intelligence Index found China is closing in on the U.S. in state-of-the-art AI. The report said that 40 AI models of note were produced by U.S.-based institutions last year, while another 15 were produced by China, and three by Europe. The report said Chinese models are closing the gap in quality, achieving near parity with U.S. models on two key benchmarks, while China has pulled ahead of the U.S. in AI publications and patents.
[ » Read full article ]

Apple's Encryption Row with U.K. Should Not Be Secret, Court Rules

A judge sided with a coalition of civil liberties groups and news organizations in ruling that a legal scuffle between the U.K. and Apple over data privacy cannot be held in secret. The U.K. wants to be able access information secured by Apple's Advanced Data Protection (ADP) system. Such access would require Apple to create a "backdoor," which it said it does not want to do over concerns it would eventually be exploited by hackers and criminals.
[ » Read full article ]

Taiwan Says China Uses Generative AI to Ramp Up Disinformation

Taiwan's National Security Bureau said in a report to Parliament that China is working to divide the public by using generative AI to spread disinformation. The report indicated that 500,000 "controversial messages" have been distributed so far this year, mainly on social media platforms. The bureau said Beijing is engaging in "cognitive warfare" and is seeking to "create division among our society."
[ » Read full article ]

NIST Calls Time on Older Vulnerabilities Amid Surging Disclosures

The U.S. National Institute of Standards and Technology (NIST) said all common vulnerabilities and exposures (CVEs) in its National Vulnerability Database (NVD) published before Jan. 1, 2018, will be marked in the NVD dataset as "deferred," and will no longer provide updates on them. NIST said, "We are assigning this status to older CVEs to indicate that we do not plan to prioritize updating NVD enrichment or initial NVD enrichment data due to the CVE's age."
[ » Read full article ]

Bringing Personal Items to Mixed Reality

Massachusetts Institute of Technology researchers developed a tool that lets users produce a digital twin of a physical item and capture its interactive features in a mixed-reality environment. The researchers created an iPhone app that lets users scan the item to generate a 3D model. Once the model is imported in the InteRecon tool interface, accessible via a mixed-reality headset, users can choose which aspects to make interactive.
[ » Read full article ]

Bacteria-Inspired Robot Roams Underwater

A bacteria-inspired underwater drone developed by researchers at UAE's Khalifa University features 12 flexible arms capable of controlled movements in any direction. Inspired by bacteria's flagella, the slender threadlike structure that allows it to swim, each arm has a motor at the base controlled by an onboard Raspberry Pi computer.
[ » Read full article *May Require Paid Registration ]

Researchers Investigate AI Threats in Software Development

Researchers led by University of Texas at San Antonio computer science doctoral student Joe Spracklen analyzed the security risks associated with package hallucinations, in which large language models (LLMs) generate code that links to a third-party software library that does not exist. This would enable a hacker to create a new package with the same name as the hallucinated package and inject malicious code. The researchers found that open-source LLMs are four times more likely than GPT-series models to produce package hallucinations, and JavaScript is more susceptible to hallucinations than Python.
[ » Read full article ]