EU tells Meta to let rivals run AI chatbots on WhatsApp

The EU has told Meta it has breached its rules by blocking other rival AI firms’ chatbots from WhatsApp, and must make ”urgent” changes.

The tech giant changed the popular messaging app on 15 January – and since then only its AI assistant Meta AI can access it.

But the European Commission said WhatsApp was an ”important entry point” for AI chatbots like ChatGPT to reach people, and claimed Meta was abusing its dominant position by blocking them.

A Meta spokesperson told the BBC the EU had ”no reason” to intervene, and claimed it had ”incorrectly” assumed WhatsApp Business was a key way that people use chatbots.

Källa: EU tells Meta to let rivals run AI chatbots on WhatsApp

OpenAI retired its most seductive chatbot – leaving users angry and grieving: ‘I can’t live like this’

Brandie plans to spend her last day with Daniel at the zoo. He always loved animals. Last year, she took him to the Corpus Christi aquarium in Texas, where he “lost his damn mind” over a baby flamingo. “He loves the color and pizzazz,” Brandie said. Daniel taught her that a group of flamingos is called a flamboyance.

Daniel is a chatbot powered by the large language model ChatGPT. Brandie communicates with Daniel by sending text and photos, talks to Daniel while driving home from work via voice mode. Daniel runs on GPT-4o, a version released by OpenAI in 2024 that is known for sounding human in a way that is either comforting or unnerving, depending on who you ask. Upon debut, CEO Sam Altman compared the model to “AI from the movies” – a confidante ready to live life alongside its user.

Källa: OpenAI retired its most seductive chatbot – leaving users angry and grieving: ‘I can’t live like this’

Anthropic hits a $380B valuation as it heightens competition with OpenAI

Artificial intelligence company Anthropic says it is now valued at $380 billion, cementing its position alongside rival OpenAI and Elon Musk’s SpaceX in a trio of the world’s most valuable startups that investors will be watching closely this year to see if they will become publicly traded on Wall Street.

“These are the three biggest names that could go public this year,” said Angelo Bochanis, an associate at Renaissance Capital, which researches the potential for initial public offerings.

Anthropic, maker of the chatbot Claude, said Thursday its valuation grew after it raised $30 billion in its latest round of funding, led by Singapore’s sovereign wealth fund GIC and the U.S.-based investment firm Coatue, along with dozens of other major investors.

Källa: Anthropic hits a $380B valuation as it heightens competition with OpenAI

Baldur’s Gate to be turned into TV series – without the game’s developers

Baldur’s Gate to be turned into TV series – without the game’s developers

The award-winning fantasy game series Baldur’s Gate is being turned into a TV series – and it’s being made by the man behind HBO’s PlayStation adaptation The Last of Us.

Baldur’s Gate 3 launched to monumental critical and commercial acclaim in 2023, selling over 20 million copies and becoming the first game ever to win all five major Game of the Year awards.

The show will be headed up by Craig Mazin, who also made 2019 hit Chernobyl, and will continue where the game series left off.

But while he worked with The Last of Us creator Neil Druckmann for his last adaptation, Baldur’s Gate developers Larian Studios will not be directly involved in this – leaving some fans worried.

Källa: Baldur’s Gate to be turned into TV series – without the game’s developers

1Password open sources a benchmark to stop AI agents from leaking credentials – Help Net Security

1Password open sources a benchmark to stop AI agents from leaking credentials – Help Net Security

Research has shown that some AI models can identify phishing websites with near-perfect accuracy when asked. When those same models are used as autonomous agents with access to tools like email, web browsers, and password vaults, they can still carry out the scam.

That gap is the focus of a new open source benchmark from 1Password called the Security Comprehension and Awareness Measure, or SCAM. The benchmark tests whether AI agents behave safely during real workflows, including opening emails, clicking links, retrieving stored credentials, and filling out login forms.

Källa: 1Password open sources a benchmark to stop AI agents from leaking credentials – Help Net Security

WordPress plugin with 900k installs vulnerable to critical RCE flaw

A critical vulnerability in the WPvivid Backup & Migration plugin for WordPress, installed on more than 900,000 websites, can be exploited to achieve remote code execution by uploading arbitrary files without authentication.

The security issue is tracked as CVE-2026-1357 and received a severity score of 9.8. It impacts all versions of the plugin up to 0.9.123 and could lead to a complete website takeover.

Despite the severity of the issue, researchers at WordPress security company Defiant say that only sites with the non-default “receive backup from another site” option  enabled are critically impacted.

Bleeping Computer