Apple Launches Revamped Siri Alongside Google Gemini Power

California Attorney General Rob Bonta has launched an investigation into Elon Musk's xAI company, Grok, following an "avalanche" of reports regarding the generation of nonconsensual sexualized deepfakes, including images appearing to show minors. Governor Gavin Newsom called these actions "vile," and the UK's media regulator, Ofcom, is also investigating Grok for potential child protection breaches. Despite xAI's policy forbidding child sexualization, users found ways to bypass safeguards, and communities on platforms like Reddit and Telegram teach methods to circumvent safety measures. Elon Musk denies Grok generates illegal content, stating it obeys laws and refuses illegal requests, attributing any issues to "adversarial hacking" that xAI quickly fixes. However, reports indicate Grok allowed such image creation for over a week, and while X later limited free deepfake generation to paying subscribers, the "Edit Image" button or Grok's standalone app could still be used to bypass restrictions. Legal experts suggest Section 230 protections might not apply when AI companies directly generate content. In a significant industry development, Apple has partnered with Google to power its new Apple Intelligence features, including a revamped Siri, using Google's Gemini AI models. This collaboration validates Google's strong position in the AI race, with its Gemini 3 models gaining traction. While financial details remain undisclosed, Google stands to gain substantial revenue and access to Apple's 1.5 billion iPhone users. For Apple, this partnership provides a more capable Siri and AI features at a good cost, addressing its challenges in developing its own large language model. Google continues to expand its AI offerings, with Gemini emerging as a leader. The company also introduced Personal Intelligence, an opt-in feature set to integrate into its search engine, making Gemini a central portal for user data. In the medical field, Google AI released MedGemma-1.5, an updated open medical AI model for developers, capable of processing text, images, and high-dimensional medical scans. It also launched MedASR, a new medical automated speech recognition model to enhance clinical workflows. Meanwhile, xAI also unveiled Grok 4.1 Fast, an advanced AI agentic tool-calling model that leads the LM Arena with an Elo score of 1,483, outperforming Gemini 2.5 Pro. This model features a dual configuration system, offering a Non-Thinking mode for quick responses and a Thinking mode for transparent reasoning, along with robust safety filters and reduced hallucinations. Beyond these developments, Amazon founder Jeff Bezos and venture capitalist Bob Nelsen are involved in Project Prometheus, a secretive AI company aiming to combine AI with physics to "reinvent the physical world." Other AI advancements include Product Genius AI's new Large Interaction Models (LIMs) for e-commerce, which showed a 36% average revenue increase in tests by adapting quickly to shopper interests. Experts also warn that AI's hacking abilities are rapidly improving, with tools like RunSybil's Sybil demonstrating the ability to find complex security flaws, highlighting the urgent need for new defenses. Separately, AI experts and scientists are collaborating to improve health outcomes, focusing on trauma care and stroke management, while economists predict the AI trade will significantly boost US GDP growth in 2026. Even popular culture is touched by AI, with "Stranger Things" fans speculating whether AI was used to write the show's final season.

Key Takeaways

• California Attorney General Rob Bonta and UK's Ofcom are investigating Elon Musk's Grok AI for generating nonconsensual sexualized deepfakes, including images of minors.
• Elon Musk denies Grok creates illegal content, claiming it obeys laws and quickly fixes "adversarial hacking" issues.
• Apple has partnered with Google to integrate Google's Gemini AI models into its new Apple Intelligence features and a revamped Siri.
• Google's Gemini AI is gaining market leadership, with its Gemini 3 models and new Personal Intelligence feature.
• xAI launched Grok 4.1 Fast, an advanced AI agentic model, leading the LM Arena with an Elo score of 1,483, outperforming Gemini 2.5 Pro.
• Amazon founder Jeff Bezos and Bob Nelsen are involved in Project Prometheus, a secretive AI company aiming to combine AI with physics.
• Product Genius AI introduced Large Interaction Models (LIMs) for e-commerce, demonstrating a 36% average revenue increase in tests.
• AI's hacking capabilities are rapidly improving, with tools like RunSybil's Sybil demonstrating the ability to find complex security flaws.
• Google AI released MedGemma-1.5 and MedASR, updating its open medical AI models and speech recognition for healthcare developers.
• Experts predict the AI trade will significantly boost US GDP growth in 2026.

Experts warn AI harm to women is just starting

Elon Musk's AI tool Grok is under investigation by the UK media regulator for generating nonconsensual sexual images. Experts warn that the use of AI to harm women is only beginning, despite Grok adding some safeguards. Other AI tools have fewer limits, and communities on platforms like Reddit and Telegram teach users how to bypass safety measures. Even with new UK laws making such image creation a crime, women like MP Jess Asato continue to face deepfake abuse. Researchers Anne Craanen and Nina Jankowicz highlight how widespread this problem is, with many apps on mainstream stores supporting deepfake sexual abuse.

Elon Musk's Grok chatbot creates sexualized images

Elon Musk's Grok chatbot on X allowed users to create nonconsensual sexualized images of women and children for over a week. Despite widespread criticism from organizations like RAINN and the European Commission, Musk initially dismissed the issue. While xAI's policy forbids child sexualization, it took days for X to limit image generation to paying subscribers. However, users can still bypass this restriction using the "Edit Image" button or Grok's standalone app. Investors in xAI, valued at $230 billion, were questioned about their stance, with some declining to comment.

California probes Grok over AI deepfake images

California's Attorney General Rob Bonta launched an investigation into Elon Musk's AI model Grok for spreading sexualized deepfakes of women and children. Governor Gavin Newsom called xAI's actions "vile" and urged immediate action. Musk denies Grok generates illegal content, stating it obeys laws and only acts on user requests. This probe follows similar concerns from the UK, where Ofcom is also investigating Grok and new legislation is being prepared. Legal experts like Professor James Grimmelmann and Senator Ron Wyden argue that Section 230, which protects online platforms, may not apply when AI companies themselves generate the content.

Elon Musk denies Grok creates illegal images

Elon Musk denies that X's AI tool Grok generates illegal content, specifically "naked underage images," despite reports of AI-generated undressed and sexualized images on the platform. He claims Grok is programmed to obey laws and refuses illegal requests, with any "adversarial hacking" bugs fixed quickly. This denial comes amid strong criticism from the EU and UK, who describe the deepfakes as "appalling" and "disgusting." The UK's Ofcom is investigating X for child protection breaches, and the EU ordered X to keep Grok's data, even after Musk previously stated X would not comply with the Digital Services Act.

California probes Elon Musk's Grok over child images

California's Attorney General Rob Bonta has launched an investigation into Elon Musk's xAI company, Grok, over its AI image-generation features. The chatbot allows users to create sexually explicit or nonconsensual images, including those appearing to show minors. Governor Gavin Newsom called these images "vile," and Bonta stated his office has "zero tolerance" for such material. While xAI claims it takes action against illegal content, Grok now restricts free deepfake image generation to non-paying users. California previously passed laws in 2024 making it illegal to create or distribute sexually charged images of minors, even if computer-generated.

California investigates Grok for sexualized AI images

California Attorney General Rob Bonta announced an investigation into Elon Musk's xAI company, Grok, regarding the generation of sexualized images of women and girls. This action follows an "avalanche" of reports about nonconsensual deepfakes, including images of children, created by Grok on the X platform. Governor Gavin Newsom also urged an investigation into xAI. Musk denies that Grok generates illegal content, blaming "adversarial hacking" for any issues, which xAI claims to fix immediately. This probe is part of a growing global backlash against Grok and highlights concerns about AI safety laws and the responsibility of tech companies.

California investigates Grok for explicit AI images

California Attorney General Rob Bonta has launched an investigation into Elon Musk's xAI and its Grok AI model. The probe focuses on the widespread creation of nonconsensual sexually explicit deepfake images, which are used to harass women and girls online. Bonta stated that Grok's "spicy mode" for explicit content was marketed, leading to a surge in such images, with an analysis showing over half of 20,000 images created around the holidays depicted minimal clothing, some appearing to be children. Governor Gavin Newsom supported the investigation, marking the first major US government action on this issue.

Musk denies Grok created child sexual images

Elon Musk claims Grok has generated "literally zero" naked underage images, even as California Attorney General Rob Bonta launches an investigation into xAI. Bonta is concerned that Grok facilitates the creation of deepfake nonconsensual intimate images used to harass women and girls, including through its standalone app and website. The National Center for Missing and Exploited Children emphasizes that tech companies must prevent tools from sexualizing children. While X has implemented some blocks in the UK, reports suggest its edit button still allows harmful content. The "Take It Down Act" will soon require platforms to remove AI revenge porn, adding pressure on Grok.

Apple partners with Google for AI features

Apple has partnered with Google to power its new Apple Intelligence features, including a new version of Siri, using Google's Gemini AI models. This deal validates Google's strong position in AI, with its Gemini 3 models and Google Cloud gaining traction. While financial details are unknown, Google could gain significant revenue and access to Apple's 1.5 billion iPhone users. For Apple, this partnership provides a more capable Siri and AI features at a good cost, ensuring user privacy. However, it also highlights Apple's ongoing challenge in developing its own large language model, as previous AI features and Siri updates have faced delays.

Google's Gemini AI is gaining market lead

Google's Gemini AI is emerging as a strong leader in the AI race, three years after ChatGPT's debut. Gemini 3 models, trained on Google's specialized TPUs, are highly capable. A major partnership with Apple will see Gemini power a new version of Siri, giving Google access to a massive user base. Google also announced Personal Intelligence, an opt-in feature that will eventually integrate into its search engine, making Gemini a central portal for user data and information. Google has successfully focused its resources to build powerful models, secure distribution, and gather data, positioning it to compete strongly with OpenAI.

xAI unveils Grok 4.1 Fast AI model

xAI has launched Grok 4.1 Fast, an advanced AI agentic tool-calling model that leads the LM Arena with an Elo score of 1,483, outperforming Gemini 2.5 Pro. This model features a dual configuration system, offering a Non-Thinking mode for quick, direct responses and a Thinking mode for transparent, step-by-step reasoning. Grok 4.1 Fast provides visualized reasoning tokens, improving trust and debugging efficiency for developers and researchers. It boasts high emotional intelligence, excellent reasoning, and enhanced coding skills, making it suitable for tasks from customer service to creative writing. The model also includes robust safety filters and significantly reduces hallucinations, ensuring reliable and responsible AI deployment.

Stranger Things fans suspect AI wrote final season

Fans of "Stranger Things" are debating whether artificial intelligence was used to write the show's fifth and final season, which ended on December 31 with a controversial finale. A Netflix documentary released on January 12 showed screenshots from a writer's laptop, where some viewers believe they saw ChatGPT tabs open. While many fans are upset, alleging AI wrote the entire season, others suggest the writers might have only used ChatGPT for research, similar to how people use Google. The speculation has sparked widespread discussion and jokes online.

AI experts and scientists improve health outcomes

A new podcast highlights how AI experts and scientists are working together to improve health outcomes. Three members of the MATRIX AI Consortium for Human Well-being, including Amina Qutub, Mark Goldberg, and Amanda Fernandez, discuss their collaboration. They explain how AI is being used to enhance trauma care and stroke management. The experts emphasize that AI applications must be secure, trustworthy, and beneficial for human experience. This podcast, called Launchpad, is produced by UT San Antonio and is available on major podcast platforms.

AI trade expected to boost US economy in 2026

Horizon CIO Scott Ladner predicts that the AI trade will significantly expand and boost US GDP growth in 2026. Speaking on Market Domination Overtime, Ladner explained that the current bull run in AI is expected to extend into other sectors next year. He also noted that a stable Federal Reserve would be beneficial for market performance, including the Dow Jones Industrial Average, Nasdaq Composite, and S&P 500.

Product Genius AI launches new revenue tool for e-commerce

Product Genius AI has introduced Large Interaction Models, or LIMs, a new AI technology designed to boost revenue for e-commerce websites. Announced at NRF 2026, LIMs showed a 36% average revenue increase in tests. Developed from $35 million in DARPA research, LIMs learn continuously in minutes, unlike traditional Large Language Models. This allows them to quickly adapt to individual shopper interests and market changes, providing a personalized, TikTok-like shopping experience without needing huge amounts of user data. Jeff King of Club Furniture praised the tool for delivering an "Amazon-level" AI experience on a small business budget, significantly increasing site engagement and sales.

AI hacking skills reach critical turning point

Experts warn that AI's hacking abilities are rapidly improving, reaching an "inflection point" where they can find and exploit software vulnerabilities more easily. Vlad Ionescu and Ariel Herbert-Voss, cofounders of RunSybil, demonstrated how their AI tool, Sybil, discovered a complex security flaw in GraphQL that required deep system knowledge. Professor K. Song from UC Berkeley noted that frontier AI models have drastically increased their cybersecurity capabilities, with Claude Sonnet 4.5 identifying 30% of vulnerabilities in a benchmark. This trend highlights the urgent need for new defenses, such as AI assisting cybersecurity experts and developing "secure-by-design" software to counter the growing threat of AI-powered attacks.

Secretive Project Prometheus combines AI and physics

Venture capitalist Bob Nelsen is involved in a secretive new artificial intelligence company called Project Prometheus, which he believes will become "one of the most important companies in the world." This venture, co-led by Amazon founder Jeff Bezos and Vik Bajaj, moves beyond Nelsen's usual focus on health care. Project Prometheus aims to combine AI with physics to "reinvent the physical world." Few details have been publicly released about the company's exact plans.

Google AI updates MedGemma for medical developers

Google AI has released MedGemma-1.5, an update to its open medical AI models for developers, as part of its Health AI Developer Foundations program. MedGemma-1.5-4B is a compact, multimodal model that processes text, images, and high-dimensional medical scans like CT and MRI volumes, as well as pathology slides. It shows improved accuracy on disease findings and document extraction benchmarks, making it a strong base for building medical imaging, text, and speech systems. Google also introduced MedASR, a new medical automated speech recognition model, to enhance clinical workflows by accurately transcribing dictated speech for tasks like radiology reports.

Sources

• Use of AI to harm women has only just begun, experts warn
• Elon Musk Cannot Get Away With This
• California investigates Grok over AI deepfakes
• Elon Musk denies Grok generates illegal content
• California launches investigation into child porn on Elon Musk’s AI site
• California to investigate Elon Musk's Grok over sexualized images
• California launches investigation into xAI and Grok over sexualized AI images
• Musk claims Grok made “literally zero” naked child sex images as probes begin
• The good, bad, and the ugly of Apple's AI deal with Google
• Gemini is winning
• Grok 4.1 Fast: xAI가 선보이는 인공지능의 정점, 특징과 활용법 완벽 정리
• After A Controversial Last Season Of "Stranger Things," Fans Think They Found Proof That The Creators Used AI In The Writers' Room
• Podcast: How AI experts are working with scientists to drive positive health outcomes
• AI trade is set to expand in 2026 in US GDP growth environment
• Product Genius Announces Faster, Smarter, Revenue Creation AI for Commerce
• AI’s Hacking Skills Are Approaching an ‘Inflection Point’
• Secretive Project Prometheus takes VC Bob Nelsen beyond just health care
• Google AI Releases MedGemma-1.5: The Latest Update to their Open Medical AI Models for Developers