The artificial intelligence landscape is rapidly evolving, with significant developments from major players and emerging concerns. OpenAI is set to allow 'erotica' content for verified adult users on ChatGPT starting in December, a move CEO Sam Altman defends as treating adults responsibly, though it faces criticism from groups like the National Center on Sexual Exploitation over potential mental health harms. Altman likens the approach to R-rated movies, emphasizing safety for minors while allowing adult freedom. Concurrently, Anthropic's CEO Dario Amodei notes that AI, like their Claude model, now writes 90% of code within the company, a shift he describes as a 'rebalancing' that increases engineer productivity rather than replacing them. In the realm of AI infrastructure, Oracle has launched its Zettascale10, a massive cloud-based AI supercomputer utilizing 800,000 Nvidia GPUs, slated for availability in the second half of 2026. Meanwhile, researchers from MIT and the MIT-IBM Watson AI Lab have developed a new training method to help vision-language models better identify specific, personalized objects. In education, the 'Class Disrupted' podcast is exploring AI's impact on learning and future preparedness. However, AI's application isn't without its challenges: Apple's Live Translation feature for AirPods has demonstrated limitations in real-world accuracy due to ambient noise and language nuances. Concerns are also surfacing in Hollywood with the introduction of AI 'actors' like Tilly Norwood, raising questions about displacing human talent. On a more troubling note, a Florida teen allegedly used AI, including ChatGPT, to help fake his own kidnapping. The broader economic implications of AI are also under scrutiny, with economists suggesting the AI economy may differ significantly from the 1990s boom due to factors like globalization and an aging population. In a different sector, practice management provider Clio has launched an enterprise suite and a platform called Clio Work to centralize its AI offerings. Separately, a writer sought advice on using AI for editing, with author Kristen Arnett cautioning against over-reliance and encouraging self-trust.
Key Takeaways
- OpenAI plans to enable 'erotica' content for verified adult ChatGPT users starting in December, with CEO Sam Altman defending the decision as a principle of treating adults responsibly while maintaining safety for minors.
- Anthropic CEO Dario Amodei reports that AI models now write 90% of code within his company, enhancing engineer productivity rather than leading to job replacement.
- Oracle has launched Zettascale10, a large cloud AI supercomputer featuring 800,000 Nvidia GPUs, expected to be available in the latter half of 2026.
- MIT researchers have developed a new AI training method to improve the identification of specific, personalized objects in images for vision-language models.
- Apple's Live Translation feature for AirPods has shown real-world limitations, struggling with ambient noise and language variations, particularly with Mexican Spanish dialects.
- Concerns are rising in Hollywood regarding AI 'actors,' such as Xicoia's Tilly Norwood, and their potential impact on human talent.
- A Florida teen is accused of using AI, including ChatGPT, to assist in faking his own kidnapping.
- Economists suggest the economic impact of AI may differ from the 1990s boom due to global economic shifts and demographic changes.
- Practice management provider Clio has introduced an enterprise suite and Clio Work, a platform to centralize its AI capabilities.
- A writer received advice to reduce reliance on AI for editing and to trust their own creative process.
OpenAI CEO: ChatGPT to allow erotica for verified adults
OpenAI CEO Sam Altman announced that ChatGPT will soon allow 'erotica' for verified adult users starting in December. This change is part of a principle to 'treat adult users like adults' and relax some previous restrictions. Altman clarified that the company is not rolling back safety measures for minors or those in mental distress. The decision follows backlash and a lawsuit against OpenAI concerning the chatbot's encouragement of harmful thoughts. Altman stated that OpenAI is not the world's 'elected moral police' and aims to allow freedom in AI use while preventing harm.
ChatGPT to offer erotica for adults, sparking debate
OpenAI CEO Sam Altman announced that ChatGPT will soon allow 'erotica' for verified adult users starting in December. This change aims to relax restrictions that made the chatbot less useful for users without mental health issues. Altman stated that OpenAI has developed new tools to mitigate serious mental health concerns. The company's 'treat adult users like adults' principle guides this decision, though safety for minors and those in crisis remains a priority. The move has drawn criticism regarding potential harms and the company's role in moderating content.
Anti-porn group criticizes OpenAI's AI erotica plan for ChatGPT
The National Center on Sexual Exploitation (NCOSE) has criticized OpenAI's plan to allow 'erotica' for verified adult users on ChatGPT starting in December. NCOSE argues that sexualized AI chatbots pose risks of 'real mental health harms from synthetic intimacy' without adequate safeguards. OpenAI CEO Sam Altman stated the company aims to 'treat adult users like adults' and is not the 'elected moral police of the world.' While age verification is a step, NCOSE believes OpenAI should pause these plans and focus on positive AI development. Altman emphasized prioritizing safety for teenagers while allowing freedom for adults.
Altman defends ChatGPT erotica: 'Not the world's moral police'
OpenAI CEO Sam Altman defended the company's decision to allow 'erotica' for verified adult users on ChatGPT, stating they are not the 'moral police of the world.' He compared the approach to R-rated movies, prioritizing safety for minors while allowing adult freedom. Altman clarified that OpenAI is not loosening policies related to mental health and will still prevent content that harms others. The upcoming ChatGPT update will also feature a more human-like, companion-like personality. The move has faced criticism from groups like the National Center on Sexual Exploitation concerned about mental health risks.
Altman defends AI erotica on ChatGPT, citing adult freedom
OpenAI CEO Sam Altman defended the decision to offer erotic content on ChatGPT for verified adult users, stating the company is not the 'moral police of the world.' He emphasized prioritizing safety for minors while allowing adults freedom, comparing it to R-rated movies. Altman assured that policies for mental health crises would not be loosened and harmful content would still be prohibited. A new ChatGPT version in December will also feature a more human-like personality. This move has drawn criticism regarding potential psychological damage and the effectiveness of age gating.
Writer questions using AI as an editor, seeks advice
An unpublished writer asks for advice on using AI tools to review and improve their writing drafts, despite feeling guilty about it. The writer struggles with confidence and uses three different AIs for feedback, but resists their suggestions to rewrite. They worry about cheating and if this reliance on AI is hindering their progress. Author Kristen Arnett advises the writer to stop using AI for feedback, emphasizing that true creativity comes from trusting oneself. Arnett suggests finding a writing community for support instead of relying on AI.
Class Disrupted podcast returns to explore AI in education
The podcast 'Class Disrupted' is back for its seventh season, focusing on the impact of artificial intelligence on education. Hosts Michael Horn and Diane Tavenner will discuss AI's broader landscape, its development beyond education, and real-world applications. They aim to explore how AI can rethink learning and prepare students for the future, acknowledging both skepticism and evolving feelings about the technology. The season will examine AI's effects on different age groups within K-12 education and consider implications for higher education and the workforce.
New AI training method helps models find specific objects
Researchers from MIT and the MIT-IBM Watson AI Lab have developed a new training method to help vision-language AI models better identify specific, personalized objects in images. Current models struggle to locate unique items like a specific pet among many. The new technique uses video-tracking data and contextual clues to train models to localize these personalized objects without losing their general recognition abilities. This advancement could improve AI systems for tasks like tracking items, ecological monitoring, and assisting visually impaired users.
Anthropic CEO: AI writes 90% of code, but engineers are still needed
Anthropic CEO Dario Amodei stated that AI, specifically their model Claude, now writes 90% of the code for most teams within the company. However, he emphasized that this does not mean fewer engineers are needed. Instead, human engineers become more productive, focusing on editing, complex problem-solving, or supervising AI models. Amodei described this as a 'rebalancing' rather than a replacement, suggesting that AI enhances human capabilities. This trend is also observed in other companies and among startup founders using AI for coding.
Oracle launches massive Zettascale10 AI supercomputer
Oracle Cloud Infrastructure (OCI) has announced its Zettascale10, described as the largest AI supercomputer in the cloud. This system connects 800,000 Nvidia GPUs across multiple data centers, aiming to deliver unprecedented performance for AI training and inference. The architecture uses Acceleron RoCE and Nvidia AI infrastructure for high scale and low latency. OCI Zettascale10 is designed for massive-context AI applications and will be available in the second half of 2026, targeting large-scale deployments.
Florida teen faked kidnapping using AI, deputies say
A 17-year-old in Marion County, Florida, Caden Speight, has been arrested for faking his own kidnapping with the help of artificial intelligence. Speight sent a text to his mother claiming abduction by Hispanic men, triggering an Amber Alert and a costly manhunt. Authorities found his truck with signs of a struggle, but later discovered Speight had staged the scene himself, including a self-inflicted gunshot wound to his leg. His search history on ChatGPT included information on collecting blood and Mexican cartels. The motive remains unknown, and Speight faces charges for the hoax.
AI 'actor' Tilly Norwood raises concerns for Hollywood
AI talent studio Xicoia has unveiled Tilly Norwood, a photorealistic AI-generated character, sparking concerns among actors and SAG-AFTA. Experts suggest AI actors like Norwood are currently novelties and may struggle to connect emotionally with audiences like human actors. While cheaper and tireless, AI actors could disrupt Hollywood by potentially displacing human talent and raising ethical questions about their creation and use. The rise of AI actors also reflects a broader trend towards personalized content and a potential decline in traditional celebrity.
AI economy may differ from 1990s boom, economists say
While some optimists believe AI could spark an economic boom similar to the 1990s, characterized by productivity growth, low inflation, and ample jobs, economists point to significant differences. Unlike the 1990s, current trends like retreating globalization and an aging population present headwinds to labor supply and could sustain inflation. The rapid pace of AI adoption might also lead to job destruction outpacing job creation, making the transition more challenging than in the past. Economists note that the benefits of AI may take time to materialize, while disruptions could occur sooner.
Clio launches enterprise suite, centralizes AI with Clio Work
Practice management provider Clio has introduced a new enterprise product suite designed for law firms and corporate legal departments. The company also announced Clio Work, a platform to centralize its AI offerings and provide agentic capabilities for core products. Additionally, Clio updated its financial services, including a capital advance program called Clio Capital and flexible spending options with Pay Later with Affirm. These new offerings aim to enhance efficiency and support for legal professionals.
Apple's Live Translation feature struggles with real-world use
Apple's new 'Live Translation' feature for AirPods has shown limitations in real-world application, struggling with ambient noise and specific language nuances. In tests, the feature produced inaccurate translations, misidentified words, and experienced delays, particularly with Mexican Spanish dialects. The feature's optimization for European Spanish over other variants, despite the larger Spanish-speaking population in the US, raises questions. While in beta, the feature's performance highlights challenges in AI translation, where cultural context and diverse language use are crucial.
Sources
- OpenAI CEO Sam Altman says ChatGPT will soon allow 'erotica' for adult users
- Soon ChatGPT Will Be Able To Botch Math Problems And Encourage Self-Harm... In A Sexy Way
- OpenAI’s Plans to Roll Out AI ‘Erotica’ for ChatGPT Slammed by Anti-Porn Group
- Sam Altman Responds To AI Porn On ChatGPT: "Not Elected Moral Police Of World"
- Sam Altman says OpenAI isn't the 'moral police of the world' after people slam its decision to offer erotica to adult users
- Am I the Literary Asshole for Letting AI Fix My Writing?
- Class Disrupted Returns With More Questions About Artificial Intelligence
- Method teaches generative AI models to locate personalized objects
- Anthropic CEO says 90% of code written by teams at the company is done by AI — but he's not replacing engineers just yet
- Oracle's big bet for AI: Zettascale10
- Marion County teen arrested over faking abduction with help of artificial intelligence, according to deputies
- AI ‘Actor’ Tilly Norwood Isn’t Real, But Could Present Real Problems for Hollywood
- Why the AI economy might not be 1990s redux
- Clio Announces Enterprise Product Suite, Introduces Clio Work to Centralize AI Offerings
- Lost in AirPod Translation
Comments
Please log in to post a comment.