OpenAI Sora 2 Policy, Sean Astin Voices AI Concerns

The artificial intelligence landscape is rapidly evolving, with significant developments impacting creative industries, financial markets, and academic institutions. OpenAI is facing scrutiny over deepfake concerns, particularly regarding its Sora 2 video platform. Actor Bryan Cranston and the SAG-AFTRA union, represented by president Sean Astin, have voiced these issues, leading OpenAI to strengthen its policies to require explicit opt-in for the use of an individual's voice and likeness. This move aligns with Hollywood's growing demand for consent and protections for artists' intellectual property, with agencies like CAA and UTA supporting the proposed federal 'NO FAKES' Act. Meanwhile, financial analyst Edward Dowd warns of a potential AI-driven stock market bubble fueled by high leverage and margin debt, advising caution with AI stocks and leveraged assets. In the crypto space, TryCoinPilot has launched AI copy-trading, incorporating models like DeepSeek and Grok, and reporting profits for early users. Academic institutions are also adapting, with Purdue University planning to integrate AI competency across all undergraduate programs. On the hardware front, Broadcom has introduced Thor Ultra, the first 800G Ethernet Network Interface Card designed for AI workloads. California has enacted new laws, including state AI regulations through Senate Bill 53. The increasing sophistication of AI is also evident in the growth of ransomware-as-a-service groups, which are leveraging AI for automation and enhanced capabilities. Journalists are finding side income by training AI, though facing ethical dilemmas and declining pay. Mercer Investments is exploring AI to enhance client investment capabilities through a 'human in the loop' approach, focusing on responsible development.

Key Takeaways

• OpenAI is enhancing its policies for Sora 2 to require opt-in for voice and likeness usage following concerns from Bryan Cranston and SAG-AFTRA.
• SAG-AFTRA president Sean Astin and major talent agencies support the proposed federal 'NO FAKES' Act to protect performers from AI misuse.
• Financial analyst Edward Dowd is cautioning about a potential AI-driven stock market bubble due to high leverage and margin debt.
• TryCoinPilot has launched AI copy-trading, featuring models like DeepSeek and Grok, with early users reporting profits.
• Purdue University will integrate AI competency into all undergraduate programs to prepare students for the job market.
• Broadcom has released Thor Ultra, the industry's first 800G Ethernet Network Interface Card for AI workloads.
• California has enacted new state AI regulations via Senate Bill 53.
• Ransomware-as-a-service groups are increasingly using AI-powered tools for automation and growth.
• Journalists are training AI for supplemental income but face ethical concerns and decreasing pay.
• Mercer Investments is using AI with a 'human in the loop' approach to improve investment capabilities and client outcomes.

Bryan Cranston, SAG-AFTRA and OpenAI Address Deepfake Concerns

Actor Bryan Cranston and the SAG-AFTRA union have confirmed that OpenAI is taking their concerns about deepfakes seriously. Cranston's likeness appeared in videos on OpenAI's Sora app without his consent, which OpenAI has called an "unintentional generation" and expressed regret for. OpenAI has strengthened its policies to require opt-in for the use of an individual's voice and likeness in AI-generated content. SAG-AFTRA president Sean Astin highlighted the need for laws like the proposed NO FAKES Act to protect performers from such misuse of technology.

Bryan Cranston Thanks OpenAI for Addressing Sora 2 Deepfake Issues

Actor Bryan Cranston expressed gratitude to OpenAI for improving its safeguards against deepfakes on its Sora 2 video platform. Users had generated his likeness without permission, prompting concerns from Cranston and the SAG-AFTRA union. OpenAI stated that while its policy requires opt-in, they regret the "unintentional generations" and have enhanced measures to prevent unauthorized use of voice and likeness. Major talent agencies like CAA and UTA co-signed a statement with OpenAI and SAG-AFTRA, emphasizing the right of performers to control their digital representation. Sean Astin of SAG-AFTRA stressed the importance of opt-in protocols and the proposed NO FAKES Act for performer protection.

Hollywood Demands Consent for AI Video Generation by OpenAI's Sora 2

OpenAI has introduced new policies for its Sora 2 AI video tool following concerns from Hollywood studios, unions, and talent agencies. The tool can create realistic videos from text prompts, leading to unauthorized AI recreations of figures like Robin Williams and Martin Luther King Jr. Zelda Williams criticized the use of her father's likeness, calling it exploitative. Major talent agencies like United Talent Agency and Creative Artists Agency are pushing for consent and protections for artists' intellectual property and likeness. Both agencies and SAG-AFTRA now support the federal 'NO FAKES' Act to regulate AI usage.

Edward Dowd Warns of AI Bubble Fueled by High Leverage and Margin Debt

Financial analyst Edward Dowd is warning about a potential AI-driven stock market bubble, citing rising leverage and record margin debt. He notes that while many acknowledge the bubble, they believe it's still early, leading to a 'greater-fool' dynamic where late investors hope to pass risk to others. Dowd suggests this elevated leverage makes markets vulnerable to sudden downturns. Traders are advised to tighten risk limits on AI stocks and leveraged assets, and monitor margin debt levels as a key indicator of potential volatility across all risk assets.

TryCoinPilot Launches AI Copy-Trading with DeepSeek and Grok Models

TryCoinPilot has launched live AI copy-trading, allowing users to follow six large language models (LLMs) in a three-step process. The platform is reportedly the first to offer LLM live copy-trading and has seen a significant increase in new users. Early adopters following the DeepSeek and Grok strategies have reported profits. TryCoinPilot is focusing on adding more AI features to help users find trading signals, market opportunities, and optimal entry points, aiming to enhance crypto trading with AI-driven strategies.

Purdue University to Integrate AI Competency Across All Undergraduate Programs

Purdue University's Senate has approved a plan to integrate AI competency into all undergraduate programs, ensuring graduates possess essential AI literacy skills. This initiative aims to prepare students for a job market where AI proficiency is increasingly required. The university will embed AI education within existing courses rather than adding new ones, focusing on practical skills, ethical implications, and critical thinking about AI's societal impact. Faculty training and course redesign will support this curriculum update, which will undergo further development by university committees.

Broadcom Unveils Thor Ultra, First 800G AI Ethernet NIC

Broadcom has introduced Thor Ultra, the industry's first 800G Ethernet Network Interface Card (NIC) compliant with UEC standards, designed for AI workloads. This new NIC can connect hundreds of thousands of XPUs to manage AI models with trillions of parameters. Thor Ultra aims to accelerate AI model training and inference, meeting the high bandwidth and low latency demands of machine learning. Its UEC compliance ensures easy integration into existing Ethernet systems, further strengthening Broadcom's AI solution portfolio.

California Enacts New Laws on AI, Immigration, Wildfires, and More

Governor Gavin Newsom has signed numerous bills into law in California, addressing key issues like immigration, wildfire recovery, and insurance. New regulations include barring most law enforcement officers from wearing masks on duty and providing protections for students during immigration enforcement. The state is also enhancing wildfire rebuilding efforts with new legislation for disaster survivors. Other signed bills focus on capping insulin prices at $35, implementing new regulations for pet sellers, criminalizing threats against schools and places of worship, and establishing state AI regulations through Senate Bill 53. Additionally, new laws aim to enhance online safety for children, provide privacy controls for social media users, allow more housing construction near transit, and ban ultra-processed foods in school lunches.

Agentic AI Projects: Understanding Failure and Ensuring Success

This article from Harvard Business Review discusses why agentic AI projects often fail and provides guidance on how to set them up for success. Agentic AI represents a significant advancement beyond simple automation or generative AI. The piece aims to help readers navigate the complexities of implementing these advanced AI systems effectively.

AI Automation Boosts Ransomware-as-a-Service Group Growth

A new report indicates that ransomware-as-a-service (RaaS) groups are increasingly using AI-powered tools to enhance automation and customization, leading to faster growth. The security firm found that the time attackers take to compromise additional devices has dropped significantly. RaaS groups are offering AI features like antivirus evasion and tools to disable security software. Groups like LockBit are embracing this model to attract more affiliates, while others like Medusa are declining due to a lack of such advanced capabilities. Currently, less than half of RaaS groups offer AI-powered features.

Non-AI Stocks Lag as Market Focuses on AI-Driven Growth

Wall Street indexes are near all-time highs, largely driven by optimism in AI-related stocks, while companies without an AI connection are underperforming. Sectors like communications services, technology, and utilities are leading gains due to AI giants and increased data center demand. Conversely, industrial stocks tied to the business cycle, advertisers, and commodity chemicals are lagging. BCA Research suggests that while the overall market may continue to rise, the underperformance of cyclical and consumer stocks indicates potential economic shifts, prompting a revision of recession forecasts.

Journalists Train AI for Side Income, Facing Ethical and Pay Concerns

Many journalists are taking on gig work training AI software, such as evaluating large language model outputs, to supplement their income. While initially offering decent pay, the work has become emotionally taxing and the compensation has decreased significantly, with some reporting rates as low as $10 per hour for tasks that require extensive unpaid preparation. Some journalists are tasked with testing AI for harmful or illegal outputs, raising ethical concerns. Despite these challenges, the experience provides valuable insights into AI technology.

Mercer Investments Uses AI to Enhance Client Investment Capabilities

Mercer Investments is exploring the use of artificial intelligence to augment its investment capabilities and provide enhanced client outcomes. The firm emphasizes a 'human in the loop' approach, using AI to analyze vast datasets and uncover insights rather than fully automating processes. Two concepts being explored are Mercer Rate Sense, to accelerate investment manager research, and Mercer Market Sense, to identify capital allocation flows. Mercer is committed to responsible AI development with strong governance, data stewardship, bias mitigation, and model transparency to ensure ethical and compliant use.

Sources

• Bryan Cranston and SAG-AFTRA say OpenAI is taking their deepfake concerns seriously
• Bryan Cranston thanks OpenAI for cracking down on Sora 2 deepfakes
• Hollywood pushes OpenAI for consent
• AI Bubble Warning: Edward Dowd Flags Rising Leverage and Record Margin Debt Risk — 3 Trading Takeaways
• TryCoinPilot Launches 6 LLM Copy-Trading Strategies: DeepSeek and Grok Users Report Profits in Live AI Crypto Trading
• Purdue University Senate moves to integrate AI competency into all undergraduate programs, addresses prescription medication changes
• Broadcom Introduces Thor Ultra, the Industry’s First 800G UEC-Compliant AI Ethernet NIC
• New California laws tackle AI, immigration, wildfires and more. Here's what to know about bills signed by Newsom
• Why Agentic AI Projects Fail—and How to Set Yours Up for Success
• AI-fueled automation helps ransomware-as-a-service groups stand out from the crowd
• What about the non-AI stocks?
• From newsrooms to AI side hustles: Why journalists are training the machines that may replace them
• Augmenting our investment capabilities – how we are using AI