October 7, 2025

news  
OpenAI Sora Launch, Sam Altman, AMD Deal

OpenAI is making waves with its new video generation app, Sora, which creates 10-second videos from text prompts. While CEO Sam Altman touts it as an 'imagination engine,' critics have labeled it a potential source of 'AI slop' or 'deepfakes.' To address concerns about misuse, Sora is introducing controls for AI self-representation and working on improved watermarking. The company also acquired personalized finance app Roi, signaling a move toward integrating personal AI companions into its offerings. At its DevDay event, OpenAI highlighted its need for substantial computing power, announcing a significant infrastructure deal with AMD. Meanwhile, the broader impact of AI in the workplace is becoming clearer, with research indicating the rise of 'workslop' – low-value AI-generated content that reduces productivity. A Stanford and BetterUp study found that employees are split into 'pilots' who use AI effectively and 'passengers' who rely on it too heavily, with the former being more optimistic and productive. This 'workslop' phenomenon, where tasks are offloaded to coworkers, can take hours to resolve and lead to resentment. In the realm of cybersecurity, Palo Alto Networks is empowering its APAC partners with AI security tools like Prisma AIRS to combat AI-driven attacks. The fight against AI misuse is also extending to voice deepfakes, with new AI technologies emerging to detect and block fraudulent calls. Retail security is also being transformed by AI and biometrics, enhancing theft prevention and staff protection. In the financial sector, Moomoo reports that Malaysian investors are increasingly using AI tools for investment strategies, with AI expected to become a core driver. Meta is also adapting, planning to use data from user interactions with its AI products, including conversations and voice data, to improve its advertising and content recommendations starting December 16, 2025. Elsewhere, AI cloud provider CoreWeave is set to acquire Monolith AI to expand its industrial AI capabilities, aiming to offer a full-stack solution for manufacturing sectors.

Key Takeaways

  • OpenAI has launched its Sora video generation app, which creates 10-second videos from text prompts, and is adding controls for AI self-representation to mitigate misuse concerns.
  • Concerns exist about AI-generated content, with terms like 'AI slop' and 'deepfakes' being used, alongside discussions on the potential for AI to create novel artistic and imaginative videos.
  • Research indicates that AI is creating 'workslop' in US workplaces, reducing productivity as low-value AI-generated content makes tasks harder for coworkers, with each incident taking about two hours to resolve.
  • Employees are divided into 'pilots' who use AI to enhance their work and 'passengers' who rely on it for tasks, with 'pilots' showing higher optimism and productivity.
  • Palo Alto Networks is equipping its APAC partners with AI cybersecurity tools, including the Prisma AIRS platform, to combat the rise of AI-driven cyber-attacks.
  • New AI technologies are being developed to combat voice deepfakes, aiming to detect and block fraudulent calls and protect against AI-driven fraud.
  • AI and biometrics are enhancing retail security through improved theft prevention, staff protection, and checkout efficiency.
  • OpenAI acquired personalized finance app Roi, intending to integrate its AI companions into future products and make investing more accessible.
  • OpenAI announced a significant infrastructure deal with AMD to support its substantial computing power needs.
  • Meta plans to use data from user interactions with its AI products, including conversations and voice data, for advertising and content recommendations starting December 16, 2025.

Sora adds controls for AI video self-representation

OpenAI's Sora app now allows users to control how AI-generated versions of themselves appear in videos. These new features aim to manage user concerns about deepfakes and inappropriate content. Users can set restrictions, such as preventing their AI self from appearing in political videos or using certain language. The app is also working on improving its watermark feature to prevent misuse. These updates come as Sora faces criticism for generating low-quality content.

OpenAI launches Sora video app amid 'AI slop' debate

OpenAI has launched its new video generation app, Sora, which creates 10-second videos from text prompts. CEO Sam Altman calls it a powerful 'imagination engine,' while critics label it a 'TikTok for deepfakes' or an 'avalanche of AI slop.' The app is currently available to a select group of users. Despite its popularity, concerns remain about the potential for misuse and the quality of AI-generated content.

Sora app showcases AI's creative potential beyond 'slop'

OpenAI's new video app, Sora, offers a glimpse into the creative possibilities of artificial intelligence. The app generates imaginative videos, from a judo match between a woman and an elephant to a figure skater on Saturn. It even features a humorous clip of OpenAI's founder, Sam Altman, shoplifting. While some dismiss AI-generated content as 'slop,' Sora demonstrates its potential for artistic and novel creations.

AI 'workslop' creates two employee types: pilots and passengers

The rise of AI in the workplace has led to two types of employees: 'pilots' who use AI to enhance their work and 'passengers' who rely on it to do tasks for them, creating 'workslop.' This low-value, AI-generated content floods inboxes and reduces productivity, costing companies significantly. Research from Stanford and BetterUp highlights that 'pilots' are more optimistic and productive. Companies are advised to guide AI usage to improve efficiency rather than simply adopting it as a shortcut.

AI 'workslop' reduces productivity in US workplaces

New research indicates that artificial intelligence is decreasing productivity in American workplaces due to the creation of 'workslop.' This low-effort AI-generated content, which appears polished but lacks substance, makes tasks harder for coworkers. A survey found 40% of employees received workslop last month, with each incident taking about two hours to resolve. This phenomenon, described as 'cognitive offloading' to other humans, can lead to resentment and a decline in trust among colleagues.

Palo Alto Networks empowers APAC partners with AI cybersecurity

Palo Alto Networks is equipping its partners in the Asia-Pacific region with tools and training to address the growing demand for AI in cybersecurity. The company recently launched Prisma AIRS, an AI security platform designed to protect the entire AI ecosystem. With cyber-attacks becoming AI-driven, Palo Alto Networks emphasizes the need for rapid detection and remediation. Their partners are crucial in delivering these advanced AI security solutions to customers navigating the complexities of AI adoption.

AI combats AI: New tech fights voice deepfakes

The fight against AI-generated deepfakes is escalating, with new technology emerging to combat voice cloning. Companies are developing AI tools that can detect and block fraudulent voice calls, protecting banks and individuals from AI-driven fraud. This special series on The Indicator from Planet Money explores how AI is being used to counter its own malicious applications, particularly in the evolving landscape of digital crime.

AI and biometrics transform retail security

Artificial intelligence and biometrics are revolutionizing retail security by enhancing theft prevention, staff protection, and checkout efficiency. AI-powered systems can identify known offenders, detect suspicious behavior, and even track stolen items. Biometrics like facial recognition and fingerprint scanning are used for access control and employee monitoring. These technologies offer advanced capabilities to identify threats and protect assets, making retail environments safer and more secure.

OpenAI DevDay: Altman discusses AI future with Jony Ive

OpenAI held its DevDay event in San Francisco, featuring a keynote by CEO Sam Altman and a discussion with former Apple design chief Jony Ive. The event showcased advancements from OpenAI's research, product, and engineering teams, with over 1,500 developers in attendance. OpenAI also announced a significant infrastructure deal with AMD, highlighting the company's need for substantial computing power. The conversation with Ive focused on thoughtful and responsible AI development.

OpenAI acquires personalized finance app Roi

OpenAI has acquired Roi, a startup that developed an AI-powered app for personalized finance management. Roi provided tailored insights and guidance through AI companions, aiming to make investing more accessible. The acquisition aligns with OpenAI's goal to integrate personalization into its future AI products. Roi's team will join OpenAI, and its operations will cease on October 15, 2025, as they contribute to building adaptive and personal AI companions.

CoreWeave to acquire Monolith for industrial AI expansion

AI cloud provider CoreWeave has agreed to acquire Monolith AI, a company specializing in applying AI to physics and engineering challenges. This acquisition aims to combine Monolith's simulation and machine learning capabilities with CoreWeave's AI cloud platform. The goal is to offer a full-stack solution for industrial and manufacturing sectors, accelerating R&D cycles and product development. Monolith's technology is already used by companies like Nissan and BMW to shorten development times.

Moomoo sees Malaysian investors increasingly using AI tools

Futu's Moomoo platform observes Malaysian investors relying more on AI tools for investment strategies. CEO Ivan Mok expects AI to become a core driver, moving beyond a supplemental role. The platform's AI Investment Assistant and analytics help users simulate scenarios and identify opportunities. Moomoo differentiates itself by offering professional-grade tools, market data, and community features, making investing more accessible and empowering for local investors.

Meta to use AI chat data for ads

Meta will update its privacy policy on December 16, 2025, to include data from user interactions with its AI products for advertising and content recommendations. This means conversations with Meta AI chatbots, as well as voice and image data, will be incorporated into its systems. This change aims to enhance personalization in Meta's services by leveraging user engagement with its AI features.

Sources

OpenAI Sora AI video generation Deepfakes AI ethics AI content moderation AI in cybersecurity AI voice cloning AI in retail security Biometrics AI in finance AI in manufacturing AI in investment AI in advertising Meta AI AI productivity Workslop AI development AI infrastructure Personalized AI