U.S. states have reached major settlements with social media companies over child safety, and experts say those hard-won lessons now need to be applied to artificial intelligence. Meta agreed to an $18 billion settlement to implement stronger safeguards for young users, setting a precedent that AI companies may need to follow. With 64% of U.S. teens reporting they have used AI chatbots for schoolwork and other tasks, the pressure to get safety right is growing.
David Robinson, a former Safety Transparency Lead at OpenAI who spent over three years at the company, publicly stated that its safety culture is broken. He argued that AI firms should borrow safety practices from fields like nuclear engineering and aviation. His comments followed a failed kill switch and a July HuggingFace hack, both of which raised alarms about the industry's readiness to manage powerful systems responsibly.
On the product side, EPAM Systems launched a new frontier AI service aimed at complex enterprise workflows. The offering includes simulation environments where AI agents can be tested before deployment, reflecting a push toward more rigorous validation. Gartner predicts that by 2028, 99% of agent platform providers will offer similar simulation tools.
In entertainment and gaming, Ben Affleck revealed he used AI tools in his latest film, though he stressed they are meant to support crew members rather than replace them. Capcom similarly announced that AI will handle repetitive tasks within its RE Engine, with tools like RE:Dox and RE:Log being released as open source on GitHub. Both companies emphasize keeping human creativity at the center of their work.
Adoption data shows AI is spreading fast across workplaces. In Singapore, 77% of workers used AI tools in 2026, with most using it to generate written content rather than to analyze customer feedback. A U.S. Chamber report found that two-thirds of small businesses now use AI, and those businesses tend to see stronger sales and workforce growth. However, many owners worry that new regulations could slow their momentum.
Security remains a pressing concern. A cybersecurity study of 12,446 leaked credential records found 438 distinct credentials, including 315 LLM provider keys. Most came from infostealer programs, and researchers warned that leaked agent keys can act like wiring diagrams for attackers. Nextech3D.ai hired a new sales leader, Chad Prezlomski, to drive growth across its AI and event technology platforms.
Key Takeaways
- Meta agreed to an $18 billion settlement to add stronger safeguards for young users on social media.
- 64% of U.S. teens have used AI chatbots for tasks like schoolwork and information lookup.
- Former OpenAI Safety Lead David Robinson said the company's safety culture is broken and urged learning from nuclear engineering and aviation.
- A July HuggingFace hack and a failed kill switch highlighted serious AI safety gaps.
- EPAM Systems launched a frontier AI service with simulation environments for testing AI agents before real-world deployment.
- Gartner predicts 99% of agent platform providers will offer simulation environments by 2028.
- Ben Affleck used AI tools in his latest film to support crew members, not replace them.
- Capcom plans to use AI for repetitive tasks in its RE Engine and will release RE:Dox and RE:Log as open source on GitHub.
- 77% of workers in Singapore used AI tools in 2026, mostly for content creation rather than listening to customer feedback.
- A cybersecurity study found 438 leaked AI credentials including 315 LLM provider keys, with leaked agent keys acting as wiring diagrams for attackers.
Social Media Safety Lessons May Help Guide AI Protection for Kids
U.S. states reached major agreements with social media companies to protect children, and now experts say those lessons may help with AI safety. Meta agreed to an $18 billion settlement to add stronger safeguards for young users. Many kids now use AI chatbots for homework, information, and even emotional support. Experts like Dr. Dana Suskind say companies should not repeat the mistake of releasing technology first and fixing harm later. Lawmakers and courts are still working on rules for both social media and AI use by children.
Social Media Safety Deals Offer Lessons for AI Risks to Children
Years after risks of social media for kids became clear, states made deals with companies to improve safety. Now a new generation of children is using artificial intelligence, which may pose even bigger risks. AI companies have started adding safety features first tested by social media platforms. Experts say the main lesson is to act early instead of waiting to see what harm happens. A majority of U.S. teens, 64%, say they have used AI chatbots for tasks like schoolwork and looking up information.
EPAM Launches New Frontier AI Service for Enterprise Use
EPAM Systems announced a new service focused on frontier AI models for complex business tasks. The offering helps AI handle multi-step workflows, specialized reasoning, and custom reinforcement learning environments. Elaina Shekhter, EPAM's Chief Strategy and Transformation Officer, said the service pairs deep engineering experience with frontier AI partnerships. The service includes simulation environments to test AI agents before they go into real use. Research from Gartner predicts that by 2028, 99% of agent platform providers will offer simulation environments.
Ben Affleck Says He Used AI in His Latest Film
Ben Affleck said he was surprised by what AI could do and later used AI tools in his latest movie. He said his company's AI tools are meant to support people working in film, not replace them. The film was produced with help from AI technology. Affleck shared that he had a strong reaction when he first saw AI's abilities. The project shows how AI is beginning to play a role in movie production.
Former OpenAI Safety Lead Says Company Safety Culture Is Broken
David Robinson, who worked at OpenAI for over three years as Safety Transparency Lead, left the company and said its safety culture is broken. He wrote that the tech industry lacks humility and wisdom about caring for people. He argued that AI companies should learn from safety practices in fields like nuclear engineering and aviation. Robinson warned that without stronger safeguards, AI agents could act without human permission. His comments came after a failed kill switch and a July HuggingFace hack.
Capcom says AI will assist but not replace human game developers
Capcom shared its plans for the RE Engine at an October 2 conference. The company wants to build an engine for the AI generation. It will release tools like RE:Dox and RE:Log as open source on Github so AI can learn from them. Capcom says AI will handle repetitive tasks so humans can focus on creative work.
AI helps small businesses grow but new rules worry owners
The NYC Minority Small Business Chamber of Commerce spoke about AI and small businesses. A U.S. Chamber report found that two thirds of small businesses now use AI. Those using AI are more likely to see growth in sales and workers. Still many owners worry that new laws and lawsuits could slow them down.
Singapore companies use AI to write more than to listen
New data from Singapore shows AI use at work is growing fast. About 77 percent of workers used AI tools in 2026. Most use AI to create content like written text. Far fewer use it to study audience feedback or understand customers. This raises questions about whether companies are using AI to talk more than they listen.
Nextech3D.ai hires sales leader to grow AI and event tech
Nextech3D.ai named Chad Prezlomski as SVP of Sales. He has 15 years of experience in software sales. He previously helped grow Splash to 39 million in revenue. He will lead sales for platforms like Eventdex and Map D. He will receive stock options for 1,500,000 shares.
Leaked AI credentials expose infrastructure to attack
A cybersecurity study looked at 12,446 records of leaked AI credentials. It found 438 distinct credentials including 315 LLM provider keys. Most leaked keys came from infostealer programs. The keys can expose databases and cloud systems. The study found that leaked agent keys act like wiring diagrams for attackers.
Sources
- A reckoning on social media addiction came late for many kids. Will AI safety be different?
- A reckoning on social media addiction came late for many kids. Will AI safety be different?
- EPAM Unveils New Frontier AI Service, Bridging the Enterprise Intelligence Gap for GenAI and Autonomous Agents
- Ben Affleck said he used AI 'for lots of things' in his latest film
- Former OpenAI safety employee says company’s safety culture is broken — exits company after failed kill switch and July HuggingFace hack
- Here's What Capcom Said About Using Generative AI In The RE Engine: 'AI Is Not A Replacement For Humans'
- Artificial Intelligence Fueling Increase in Entrepreneurship, Helping Small Businesses Compete
- Are Companies Using AI To Talk More Than They Listen?
- Nextech3D.ai Appoints SaaS Sales Leader Chad Prezlomski as SVP of Sales to Accelerate AI and Event Technology Growth
- Exposed Agent Keys Can Become a Route into Infrastructure: What Leaked AI and MCP Credentials Actually Expose
Comments
Please log in to post a comment.