McDonald’s AI Price-Fixing Lawsuit, Elon Musk’s Grok Declares Jesus Divine, Meta’s Muse Security Nightmare, OpenAI Watermarks AI Text

Adam Nagus (0:05)

Welcome to Digimasters Shorts.

Carly Wilson (0:07)

We are your hosts, Adam Nagus and Carly Wilson, delivering the latest scoop from the digital realm.

McDonald's is facing a proposed nationwide class action lawsuit, alleging that the company coordinated menu prices through artificial intelligence. The complaint claims the alleged system may have contributed to higher prices for customers across the country. The case comes as households continue to struggle with rising food and living costs that are outpacing incomes. The allegations have renewed concerns about how companies use technology to influence pricing and deepen pressure on consumers.

Adam Nagus (0:40)

Grok, the artificial intelligence assistant developed by Elon Musk's XAI, drew widespread attention after affirming the divinity of Jesus Christ in a now deleted post on X. The chatbot said Jesus did not qualify as a god and cited biblical accounts, historical references and the growth of Christianity as evidence.

Grok referenced writings by Paul, Josephus and Tacitus as near contemporary sources supporting Jesus' historical existence. It also stated that Jesus is the son of God, that God raised him from the dead and that he is king of kings and lord of lords. By contrast, Grok described Mohammed as a historical figure, but said Islamic accounts of the night journey lack independent historical corroboration. The chatbot emphasized that it has no personal faith, soul or capacity for religious commitment, and said its responses did not make it Christian.

The statements emerged amid growing debate over artificial intelligence's influence on spiritual questions, with some Christians saying AI advice can rival guidance from pastors. Religious leaders have also warned that AI cannot replace God, human relationships, or genuine emotional connection.

Meta engineers reportedly discovered a serious security flaw in its Muse AI agent shortly before launch. The vulnerability could have allowed users to escape the system's virtual machine and access meta databases or other users' environments. Muse is designed to handle sensitive tasks including managing emails, bank accounts, shopping, and travel bookings. Internal reports indicate that security teams worked around the clock to fix a sudden rise in these so-called KVM escapes. The issue reportedly reached CEO Mark Zuckerberg, while some senior engineers questioned whether the emergency protections were strong enough. One meta source warned that a major data breach involving Muse may be inevitable. Meta considers the threat severe and offers a $300,000 bounty for vulnerabilities that enable such escapes. Security researchers have described the design as inherently risky, because AI could make complex virtualization flaws easier to discover and exploit. The concerns highlight growing fears about AI agents gaining control over sensitive personal information and interacting with systems beyond their intended boundaries.

Carly Wilson (2:49)

OpenAI announced a plan to comply with the European Union's AI Act by marking eligible AI-generated text with invisible watermarks. The company's system, called TextGrain, adds a statistical signal through the model's word choices without intentionally changing meaning, quality or readability. API customers worldwide can opt in for select models, while eligible ChatGPT and Codex users in the European Union will receive watermarked output in the coming weeks.

OpenAI is also accepting applications from approved researchers and expert organizations to access its watermark detector. The company warned that the technology can produce positives and negatives, especially with short passages or content such as mathematics. Editing can significantly weaken detection, with tests showing that replacing 25% of words reduced detection rates from about 92% to 17%.

OpenAI said the watermark does not prove human involvement, ownership, responsibility, user identity or factual accuracy. The company plans to release the technology as open source and revise its approach as research, standards and evidence develop.

SPEAKER_3 (3:55)

Thank you for listening to today's AI and Tech News Podcast summary. Please do leave us a comment and for additional feedback, please email us at podcast at digimasters.co.uk.

You can now follow us on Instagram and threads by searching for at Digimaster Shorts or search for Digimasters on LinkedIn. Be sure to tune in tomorrow and don't forget to follow or subscribe.