Welcome to the June 16, 2025 edition of ACM TechNews, providing timely information for computer professionals three times a week.

HEADLINES AT A GLANCE

• U.K. Government Passes Data Bill
• China's First Invasive BCI Trial Achieves Positive Results
• Cyberattack on Washington Post Strikes Journalists’ Email Accounts
• Libraries Open Their Stacks to AI
• Big Tech Tests Datacenter Flexibility
• Industry Calls for More Research Funding, Public-Private Partnerships in U.S. AI Strategy
• Quantum Mechanics Tapped to Make a Factory for Random Numbers
• AI Bots Take Over the Web
• Secure Boot Under Threat
• Art Restoration Gets a High-Tech Makeover
• Software Vulnerabilities Pile Up at Government Agencies
• Google, U.S. Experts Join on AI Hurricane Forecasts

U.K. Government Passes Data Bill

The U.K.'s Data (Use and Access) Bill has cleared its final parliamentary hurdle and will become the Data (Use and Access) Act 2025 once it receives Royal Assent. Among other things, the bill facilitates patient data-sharing by National Health Service trusts, allows for the development of a 3D underground map of the nation's pipes and cables, and implements new legal offenses for the malicious use of deepfake technology. A proposed amendment that would have required tech companies to disclose the use of copyrighted materials in AI model training was rejected.
[ » Read full article ]

China's First Invasive BCI Trial Achieves Positive Results

Researchers at the Chinese Academy of Sciences' Center for Excellence in Brain Science and Intelligence Technology and Fudan University's Huashan Hospital in China implanted a brain-computer interface (BCI) in a patient with quadriplegia in March. Weeks after the surgery, the patient was able to play racing games and chess on a computer using only his mind. The procedure reportedly was China's first clinical trial of high-throughput wireless invasive BCI.
[ » Read full article ]

Cyberattack on Washington Post Strikes Journalists’ Email Accounts

The email accounts of several journalists at The Washington Post were compromised in a cyberattack potentially linked to a foreign government, officials at the media company told affected staffers. The officials said the intrusions compromised journalists’ Microsoft accounts and could have granted the intruder access to work emails they sent and received. Targets included those on the national-security and economic-policy teams, including some who write about China.
[ » Read full article *May Require Paid Registration ]

Libraries Open Their Stacks to AI

With support from Microsoft and OpenAI, Harvard University's Institutional Data Initiative has released a dataset containing more than 394 million scanned pages from nearly 1 million books in 254 languages for use by AI researchers. The Boston Public Library also made a deal with OpenAI to digitize its collection, while Google and Harvard are collaborating to retrieve and release public domain volumes from Google Books for use in AI training.
[ » Read full article ]

Big Tech Tests Datacenter Flexibility

As part of the Electric Power Research Institute's DCFlex Initiative, three datacenter hubs will test experimental methods to ensure flexibility and adaptability to local power grid needs. In Lenoir, NC, Google and Duke Energy will collaborate on workload choreography. A second site in Phoenix will see two datacenter workloads provide AI training and cloud services from Oracle and Nvidia while Emerald AI coordinates the choreography with local utilities. In Paris, France, Schneider Electric and transmission system operator RTE will test an uninterruptible power supply system.
[ » Read full article ]

Industry Calls for More Research Funding, Public-Private Partnerships in U.S. AI Strategy

In recent comments on the 2025 National AI R&D Strategic Plan, Google, Amazon, IBM, and Anthropic called for the U.S. government to prioritize investments in AI-related research and development and to lead in the creation of AI standards that could be adopted worldwide. Google also encouraged the creation of "a comprehensive initiative to educate America's science students and scientific workforce on new AI technologies, supporting the uptake of AI as the next scientific instrument."
[ » Read full article ]

Quantum Mechanics Tapped to Make a Factory for Random Numbers

The Colorado University Randomness Beacon (CURBy), developed by researchers at the U.S. National Institute of Standards and Technology (NIST) and the University of Colorado Boulder, is the first random number generator that produces verifiable random numbers using quantum entanglement. The researchers also developed the Twine protocol, a set of quantum-compatible blockchain technologies that mark each set of data for CURBy with a hash that allows them to be traced and verified.
[ » Read full article ]

AI Bots Take Over the Web

AI companies like OpenAI and Anthropic have deployed bots for real-time retrieval and recapping of content, even as more people use chatbots as an alternative to Google searches. New York startup TollBit found that retrieval bot traffic to 266 websites, with national and local news organizations accounting for half, surged 49% from the fourth quarter of 2024 to the first quarter of 2025.
[ » Read full article *May Require Paid Registration ]

Secure Boot Under Threat

Security researchers have identified two flaws in Secure Boot that enable attackers to bypass the industry-standard protocol. Microsoft has patched the first flaw, affecting a DT Research-created module used by more than 50 original equipment manufacturers' devices, which could let attackers completely skip Secure Boot. The second flaw, not yet addressed by Microsoft, involves the igel-flash-driver Linux kernel module and could enable attackers to skirt Secure Boot's bootkit protections.
[ » Read full article ]

Art Restoration Gets a High-Tech Makeover

Massachusetts Institute of Technology mechanical engineer Alex Kachkine has developed a method to repair and retouch damaged paintings that leverages advanced image processing and precision printing technologies. The process involves performing a high-resolution digital scan to map damaged areas and reconstructing the painting's original appearance using machine learning algorithms and manual adjustments in Adobe Photoshop. Restoring a Renaissance-era painting using this process took only 3.5 hours.
[ » Read full article ]

Software Vulnerabilities Pile Up at Government Agencies

A report by application security company Veracode found around 80% of government agencies have unresolved software vulnerabilities (security debt) that have not been addressed for at least a year, and 55% have even longer-standing security debt. Third-party and open-source software account for 70% of critical security debt but only 10% of overall security debt. Veracode indicated some of this security debt is associated with the use of applications built on legacy frameworks that are no longer supported by developers.
[ » Read full article ]

Google, U.S. Experts Join on AI Hurricane Forecasts

The U.S. National Hurricane Center has entered into a cooperative research and development agreement with Google's DeepMind to improve its hurricane forecasts. DeepMind's upgraded AI weather forecasting model can track a storm's development for up to 15 days and predict both its path and strength. The company said its hurricane intensity forecasts "are as accurate as, and more often accurate than," conventional forecasts.
[ » Read full article *May Require Paid Registration ]