OpenAI develops Deployment Simulation to forecast AI risks

Key Takeaways

Brands secretly use AI-generated influencers to promote products

Some brands are using AI-generated influencers to promote their products on social media without clearly labeling them as fake. This practice has raised concerns about transparency and trust. Consumer groups are calling for brands to be open about using AI-generated content. The European Union has new rules requiring AI-generated content to be labeled, but these rules don't apply in the UK. Brands like Ashle and Maket are using AI-generated influencers, but Ashle deleted their AI-generated images after being questioned.

Brands use AI-created influencers on social media

An investigation found that brands are using artificial intelligence-generated influencers to promote products on social media. These AI-generated influencers appear to be real customers but are actually computer-generated. The practice raises concerns about transparency and trust. New European Union rules require AI-generated content to be labeled, but these rules don't apply in the UK. Consumer groups are calling for brands to be open about using AI-generated content.

AI coding tools boost productivity

A study found that AI coding tools can improve productivity for software developers. The study used data from over 100,000 GitHub developers and found that each generation of AI coding tools raises coding activity more than the last. However, human bottlenecks in reviewing, integrating, testing, and deploying software remain significant. The productivity benefits of AI coding tools are largely confined to the initial stages of software development.

OpenAI method forecasts AI risks

OpenAI has developed a new method called Deployment Simulation to forecast AI risks before deployment. This method corrects for sycophantic behavior in AI models and provides a more accurate assessment of AI risks in real-life situations.

Viral AI speculative fiction sparks concern

A piece of speculative fiction about Europe in 2031 has gone viral, sparking concern about the potential risks of AI. The story suggests that Europe could be defenseless against AI-powered hackers and fall behind in the AI race. Experts warn that such speculative fiction can be misleading and should be treated with caution.

Google unveils AI products at I/O 2026

Google CEO Sundar Pichai unveiled several AI products at Google I/O 2026, including Gemini Spark, Gemini Omni, and Gemini 3.5 Flash. These products aim to improve AI's ability to simulate reality and interact with users. Google Cloud revenues grew 63% year-over-year in Q1 2026, driven in part by AI advancements.

Nvidia CEO sees AI improving lives

Nvidia CEO Jensen Huang believes that AI can improve people's lives by increasing productivity and economic growth. He argues that AI can help close the technological divide in America and create new opportunities for workers. Huang also acknowledges the need for government regulation and safety standards for AI.

Apricot STEM Fair highlights science and AI

The Apricot STEM Fair is an event that features activities and games rooted in science, technology, engineering, and math. This year's fair includes exhibits on natural science, artificial intelligence, and robotics. The event aims to promote STEM education and interest among people of all ages.

Repurposing EV batteries for AI data centers

Retired electric vehicle batteries could help solve the power problem faced by AI data centers. These batteries can be repurposed for stationary energy storage, reducing peak electricity demand and strain on the grid. Companies are exploring this approach as a cost-effective solution to meet the growing power needs of AI facilities.

Public safety AI helps keep World Cup safe

Advanced public safety AI is being used to keep the World Cup safe. The AI system fuses data from millions of devices, sensors, and cameras to provide real-time operational pictures. It enables instant incident detection, resource dispatch, and language translation. Drones are also being used to monitor crowds and assess emergencies.

OpenAI warns against government AI monopoly

A new OpenAI hire warns that a government monopoly on AI could lead to 'very scary outcomes.' This could result in a loss of innovation and progress in the field. The debate over AI regulation is complex, with different stakeholders having different views on the best approach.

Sources

• Brands using AI-generated influencers to promote products on social media
• Brands Employing AI-Created Influencers on Social Media 
• Writing code versus shipping code: Productivity effects across generations of AI coding tools
• New OpenAI Method Forecasts AI Risks Before Deployment
• Yet Another Piece of AI-Pilled Speculative Fiction Has Gone Dangerously Viral
• Sundar Pichai at I/O 2026: Gemini Spark and the Agentic Turn
• AI can improve lives, Nvidia chief says
• Apricot STEM Fair hightlights natural science, artificial intelligence
• Old EV Batteries Could Help Solve AI’s Exploding Power Problem
• A New Generation Of Public Safety AI Is Helping Keep The World Cup Safe
• New OpenAI Hire Warns Government Monopoly on AI Could Lead to ‘Very Scary Outcomes’