Welcome to the April 7, 2025 edition of ACM TechNews, providing timely information for computer professionals three times a week.

HEADLINES AT A GLANCE

• U.S. Plans to Build AI Datacenters on Federal Land
• Trump Extends TikTok Deadline for Second Time
• DARPA Picks Firms for Quantum Computer Development
• Kids Are Talking to 'AI Companions.' Lawmakers Want to Regulate That
• U.S. Weather Agency to Preserve Research Websites
• AI Bots Strain Wikimedia as Bandwidth Surges 50%
• Critical Vulnerability Found in Canon Printer Drivers
• Man Employs AI Avatar in Legal Appeal. Judge Not Amused
• Microsoft Pulls Back on Datacenters
• Companies Ignore Your Privacy Demands
• Bloomberg Has a Rocky Start with AI Summaries

U.S. Plans to Build AI Datacenters on Federal Land

Sixteen sites have been identified for the development of AI datacenters on land owned by the U.S. Department of Energy (DOE). “The global race for AI dominance is the next Manhattan project,” Secretary of Energy Chris Wright said Thursday. “With today’s action, the DOE is taking important steps to leverage our domestic resources to power the AI revolution."
[ » Read full article ]

Trump Extends TikTok Deadline for Second Time

U.S. President Trump on Friday again extended a deadline requiring China's ByteDance to sell the U.S. operations of TikTok or face an effective ban in the country. Trump said the TikTok deal “requires more work to ensure all necessary approvals are signed.” The latest extension pushes TikTok’s deadline to mid-June. ByteDance was facing an April 5 deadline to sell TikTok under a national security law signed by former U.S. President Biden last year. The original deadline was Jan. 19, but a Trump executive order gave the company 75 more days.
[ » Read full article ]

DARPA Picks Firms for Quantum Computer Development

The U.S. Defense Advanced Research Projects Agency (DARPA) has chosen the first 15 companies for the initial stage of its Quantum Benchmarking Initiative (QBI), launched in July to explore the feasibility of building an industrial quantum computer within a decade. The companies were chosen after a team of U.S. quantum experts screened all the proposers’ written abstracts and oral presentations, QBI's Joe Altepeter said.
[ » Read full article ]

Kids Are Talking to 'AI Companions.' Lawmakers Want to Regulate That

At least three bills under consideration in California aim to limit how "AI companion bots" can interact with minors. S.B. 243 would impose restrictions on addictive design features in AI companion bots, implementing protocols for handling discussions about self-harm or suicide, requiring the makers of these bots to undergo regular compliance audits, and allowing users to sue if they suffer harm due to a company's failure to comply. The other bills would ban AI companions for those age 16 and younger, and establish a statewide standards board to assess and regulate AI tools for minors.
[ » Read full article *May Require Paid Registration ]

U.S. Weather Agency to Preserve Research Websites

The U.S. National Oceanic and Atmospheric Agency (NOAA) will preserve access to websites tied to its research division, after previously moving to cancel a cloud Web services contract that supports many of its pages. The service contract for NOAA’s Office of Oceanic and Atmospheric Research was set to terminate, jeopardizing access to most of the office’s public-facing websites. “There will be no interruption in service,” the NOAA said last week. “All NOAA Research sites will remain online.”
[ » Read full article ]

AI Bots Strain Wikimedia as Bandwidth Surges 50%

The Wikimedia Foundation reported a 50% jump in bandwidth used for downloading multimedia content since January 2024, which it attributed to automated bots scraping data to train large language models. These bots account for only 35% of total pageviews, but 65% of the costliest requests to its core infrastructure. The resulting traffic surges are straining Wikimedia's Site Reliability team, code review tools, and bug trackers.
[ » Read full article ]

Critical Vulnerability Found in Canon Printer Drivers

Canon recently published an advisory that said drivers associated with several of its production printers, office multifunction printers, and laser printers were affected by an out-of-bounds vulnerability. The security hole is tracked as CVE-2025-1268 and it has a CVSS severity score of 9.4. Canon told users exploitation of the vulnerability can allow an attacker to prevent printing or potentially execute arbitrary code “when the print is processed by a malicious application."
[ » Read full article ]

Man Employs AI Avatar in Legal Appeal. Judge Not Amused

Jerome Dewald deployed an AI-trained digital avatar in a video he was allowed to show a panel of New York State judges to argue for a reversal of a lower court’s decision in his dispute with a former employer. When a judge asked whether a man in the video was his attorney, Dewald responded, “I generated that.” The judge said that she didn't "appreciate being misled,” before ordering the video turned off.
[ » Read full article *May Require Paid Registration ]

Microsoft Pulls Back on Datacenters

Microsoft has pulled back on datacenter projects worldwide, suggesting the company is wary of expanding its cloud computing infrastructure too rapidly. According to sources, the tech giant has halted talks or delayed development of sites for datacenters in the U.K., Indonesia, and Australia, as well as the U.S.
[ » Read full article *May Require Paid Registration ]

Companies Ignore Your Privacy Demands

Consumer Reports researchers found that companies may not be complying with state laws allowing residents to opt out of letting websites sell or share their personal information. The researchers used software that made it look like they were in Colorado or California, where residents can use Web browsers to complete privacy opt-out forms. Despite opting out, the researchers still saw highly targeted ads on at least 12 of 40 relatively well-known websites.
[ » Read full article *May Require Paid Registration ]

Bloomberg Has a Rocky Start with AI Summaries

Bloomberg has corrected at least three dozen AI-generated summaries of articles published this year. The summaries appear above news articles and consist of three bullet points purportedly outlining the articles' main points. Bloomberg said that "currently 99% of AI summaries meet our editorial standards," adding that the AI summaries are “meant to complement our journalism, not replace it.”
[ » Read full article *May Require Paid Registration ]