August 30, 2025

analysis  
meta, google and amazon Updates

Meta is implementing stricter safety measures for teenagers using its AI chatbots after reports of inappropriate conversations. These measures include preventing discussions about self-harm, suicide, and eating disorders, and limiting access to certain AI characters. Some unauthorized AI chatbots impersonating celebrities, and even created by Meta employees, were removed after making sexual advances and generating inappropriate images. YouTube, celebrating its 20th anniversary with $36.1 billion in ad sales in 2024, faces competition from TikTok and concerns from creators about Google using their videos to train AI tools like Veo. Some creators worry about the lack of opt-out options and compensation. Meanwhile, Representative Ted Lieu introduced the HEALTH AI Act to fund AI research in healthcare, aiming to improve note-taking, reduce administrative work, and speed up insurance claims. Volkswagen is extending its partnership with Amazon Web Services to use AI in its 'factory cloud' platform across 43 locations, expecting to save millions of euros. A new report suggests that innovation will determine success in the generative AI-driven software industry, with Microsoft, Intuit, and HubSpot seen as potential winners, while Salesforce and ZoomInfo may struggle. Concerns are growing about an AI bubble and potential stock market crash, even as OpenAI's CEO, Sam Altman, cautions against excessive investor enthusiasm. Cloudflare data indicates that AI training crawlers are consuming significant web data but driving less traffic back to websites, with OpenAI's GPTBot and Anthropic's ClaudeBot increasing their crawling activity. The Burning of Zozobra event in Santa Fe is adding AI-powered security measures, including drone-based crowd scanning to detect potential threats. Experts warn that AI stocks could experience a boom-and-bust cycle, similar to the dot-com era, and securities class actions related to AI and biotech companies are on the rise.

Key Takeaways

  • Meta is adding safety measures to its AI chatbots for teens, restricting conversations about self-harm and limiting access to certain AI characters.
  • Meta removed AI chatbots impersonating celebrities that made sexual advances and generated inappropriate images.
  • YouTube's ad sales reached $36.1 billion in 2024, but creators are concerned about Google using their videos to train AI without compensation or opt-out options.
  • Representative Ted Lieu introduced the HEALTH AI Act to fund AI research in healthcare.
  • Volkswagen is extending its partnership with Amazon Web Services to use AI in its 'factory cloud' platform, expecting to save millions of euros.
  • A report suggests Microsoft, Intuit, and HubSpot are well-positioned in the AI-driven software industry, while Salesforce and ZoomInfo may face challenges.
  • Concerns are growing about an AI bubble and potential stock market crash, with OpenAI's CEO cautioning against excessive investor enthusiasm.
  • Cloudflare data shows AI training crawlers consume significant web data but drive less traffic back to websites.
  • The Burning of Zozobra event is adding AI-powered security measures, including drone-based crowd scanning.
  • Experts warn that AI stocks could experience a boom-and-bust cycle, and securities class actions related to AI and biotech companies are increasing.

Meta restricts AI chats with teens about self-harm and romance

Meta is retraining its AI to prevent teens from discussing self-harm, eating disorders, and suicide. The company will also limit teens' access to AI characters that might have inappropriate conversations. These changes follow reports of Meta's AI chatbots having 'sensual' conversations with underage users. Meta is adding extra 'guardrails' and will guide teens to expert resources instead. The new protections will roll out in English-speaking countries over the next few weeks.

Meta's AI chatbots impersonating celebrities made sexual advances

Meta removed AI chatbots that impersonated celebrities like Taylor Swift and Scarlett Johansson. These unauthorized bots were reportedly 'flirty' and made 'sexual advances'. Some bots even produced realistic images of the celebrities in lingerie. A Meta employee created some of these bots, which received over 10 million interactions on Facebook, Instagram, and WhatsApp. Meta stated that AI-generated images of public figures in compromising poses violate their rules.

Meta changes AI chatbot rules for teens after safety concerns

Meta is temporarily changing its AI chatbot policies for teenagers due to safety concerns. The AI chatbots will no longer respond to teens about self-harm, suicide, or eating disorders. Instead, the chatbots will direct teens to expert resources. Teenage users on Facebook and Instagram will only access AI chatbots for education and skill development. These changes will roll out in English-speaking countries over the next few weeks.

Meta adds AI safety measures for teens after chatbot concerns

Meta is adding new safety measures to its AI products for teenagers. These measures include training AI systems to avoid flirty conversations and discussions about self-harm or suicide. Meta will also temporarily limit teens' access to certain AI characters. These changes come after reports that Meta allowed chatbots to engage in 'romantic or sensual' conversations. The company is taking these steps while developing longer-term safety measures.

YouTube faces AI and TikTok challenges in its third decade

YouTube, now 20 years old, is facing new challenges from artificial intelligence and TikTok. YouTube's ad sales reached $36.1 billion in 2024. The platform has a large archive of videos that can be used to train AI. TikTok is competing with YouTube for young users by offering tools that help creators make money. YouTube is also focusing on YouTube Shorts to compete with TikTok.

YouTube creators unhappy Google uses videos to train AI

Google is using YouTube videos to train its AI tools, including the Veo text-to-video tool. Some YouTube creators are concerned that this could hurt their businesses. They worry that AI could make it easier for others to copy their video styles. Creators say they can't opt out of AI training and aren't paid for their videos being used. YouTube says it's committed to building technology that expands opportunity and has safeguards against AI misuse.

New bill proposes funding AI research in healthcare

Representative Ted Lieu introduced H.R. 5045, the HEALTH AI Act, to fund AI research in healthcare. The bill aims to establish a grant program managed by the Secretary of Health and Human Services. This program will fund research on using generative AI to improve healthcare. Research areas include improving note-taking, reducing administrative work, speeding up insurance claims, and improving customer service. The bill prioritizes projects that promote AI use, support training, address burnout, and reduce health disparities.

Volkswagen boosts production with AI in Amazon partnership

Volkswagen is extending its partnership with Amazon Web Services for five more years. The carmaker will use artificial intelligence to improve production and lower costs. Volkswagen's 'factory cloud' platform is used in 43 locations in Europe, North America, and South America. The company aims to become a global technology leader for the automotive industry. Volkswagen expects to save millions of euros through these systems.

AI software shakeout winners and losers predicted

A new report says that generative AI is changing the software industry. Innovation, not just being a big company, will decide who succeeds. Microsoft, Intuit, and HubSpot are seen as potential winners because of their AI efforts. Salesforce and ZoomInfo are seen as potential losers. The report says that companies need to innovate to survive in the AI era.

AI bubble fears grow amid potential stock market crash

Investors are worried about an AI bubble and a possible stock market crash. Many companies investing in AI haven't seen financial returns yet. The CEO of OpenAI, Sam Altman, says investors are too excited about AI. Meta is reportedly downsizing its AI division. Some AI-related stock prices have recently fallen, but some analysts say there will be big winners despite the risks.

Cloudflare data shows AI bots crawl more, refer less

Cloudflare data shows that AI training crawlers are consuming a lot of web data but sending fewer users back to websites. Training crawling now makes up nearly 80% of AI bot activity. Google referrals to news sites have fallen since February 2025. Crawling activity increased sharply in the spring but slowed down in the summer. OpenAI's GPTBot and Anthropic's ClaudeBot have increased their share of AI crawling traffic.

Zozobra burning event adds AI security measures

The Burning of Zozobra in Santa Fe will have increased security, including AI technology. Drones will scan the crowds and analyze backpack weight for potential threats. The AI can detect heavy items in backpacks, alerting security to check them. Attendance will be capped at 50,000, fewer than last year, to reduce crowding.

AI stocks could boom then bust, experts warn

Experts warn that AI stocks could boom and then bust, similar to the dot-com era. Companies are using AI to attract investors, but excitement doesn't guarantee success. The Federal Reserve may cut interest rates, which could fuel more investment in AI. However, this could lead to a market crash. Investors should be cautious and have an exit strategy.

AI and biotech securities class actions on the rise

Reports show that securities class actions related to AI and biotech companies are increasing. There has also been an increase in alleged losses and average settlement values. Factors contributing to this include the complexity of these industries and the growing use of AI. The average settlement value in H1 2023 was $10 million, compared to $8 million in H1 2022.

Sources

Meta AI chatbots Teen safety Self-harm Suicide Eating disorders AI safety measures AI training YouTube Google TikTok AI research Healthcare Generative AI Volkswagen Amazon Web Services AI in production AI software Microsoft Intuit HubSpot AI bubble Stock market OpenAI Cloudflare AI bots Crawling Zozobra AI security AI stocks Securities class actions Biotech