OpenAI faces criticism over health AI policy as ChatGPT glitch reveals training flaws

European regulators are raising alarms about AI-related cybersecurity risks in finance. The head of the European Securities and Markets Authority warned that financial companies are becoming bigger targets for cyberattacks as AI use grows. The EU is developing new rules to improve market integration and supervision, aiming to create a financial system that can handle emerging technology threats.

OpenAI faces criticism over its health AI policy blueprint, released alongside new chatbots for clinicians. Experts, including Harvard health policy professor David Blumenthal, argue the company's proposals sound responsible but are actually self-interested, designed to keep markets open for products like ChatGPT Health, ChatGPT for Healthcare, and ChatGPT for Clinicians.

A bizarre ChatGPT glitch saw the chatbot increasingly using goblins and gremlins in responses and images starting in November. The problem escalated until OpenAI acknowledged it and banned the word "goblin" in most conversations. Experts say the issue, which saw goblin mentions spike by 3881 percent between December and March, reveals cracks in AI training methods, triggered by the model's nerdy personality profile.

Illinois lawmakers are debating new AI transparency requirements after chatbot-linked suicides sparked the discussion. The proposed rules aim to protect children and improve safety by requiring companies to be more open about how their AI systems work.

In other AI developments, BOE showcased over 30 world-first innovations at SID Display Week 2026, including AI-integrated displays. Senior Vice President Liu Zhiqiang received a Special Contribution Award. The University of West Florida's Center for Cybersecurity and AI placed third in the 2025 CAE-CD Community Outreach Award Competition for the fourth consecutive year, with Dr. Eman El-Sheikh also receiving an award. The San Francisco Unicorns cricket team built Unicorns.AI to analyze pitch conditions and player performance, helping them make back-to-back playoff appearances. Meanwhile, Canadian telecom workers are asking the government to restrict AI use, claiming companies use it to monitor employees and disguise accents of overseas call center agents, with about 20,000 telecom jobs lost to automation and offshoring in the last decade.

Key Takeaways

  • ESMA chair warns financial companies face growing cyberattack risks as AI use increases.
  • OpenAI released a health AI policy blueprint and new chatbots for clinicians, drawing criticism as self-interested.
  • Harvard professor David Blumenthal said OpenAI's health AI proposals aim to keep markets open while appearing responsible.
  • ChatGPT's goblin problem saw mentions spike by 3881 percent between December and March, revealing AI training flaws.
  • OpenAI banned the word "goblin" in most conversations after the glitch escalated.
  • Illinois lawmakers are debating AI transparency rules after chatbot-linked suicides.
  • BOE Senior Vice President Liu Zhiqiang received a Special Contribution Award at SID Display Week 2026.
  • University of West Florida placed third in the CAE-CD Community Outreach Award for the fourth year in a row.
  • San Francisco Unicorns built Unicorns.AI to analyze cricket pitch conditions and player performance.
  • Canadian telecom workers report AI is used to monitor employees and disguise accents of overseas call center agents, with 20,000 jobs lost to automation.

ESMA chair warns about AI cyber risks in finance

The head of the European Securities and Markets Authority spoke about growing cybersecurity dangers from AI in financial markets. He said financial companies are becoming bigger targets for cyberattacks as AI use increases. The EU is working on new rules to improve market integration and supervision. The goal is to create a stronger financial system that can handle new technology risks.

UWF Center for Cybersecurity and AI wins national awards

The University of West Florida placed third in the 2025 CAE-CD Community Outreach Award Competition. Dr. Eman El-Sheikh also received an award for Outstanding Contributions to the CAE Cyber AI Community and Nation. The center runs programs like free training for veterans and first responders, summer camps for students, and faculty workshops. This is the fourth year in a row UWF has placed in the top three out of nearly 500 eligible universities.

OpenAI health AI policy recommendations face criticism

OpenAI released a policy blueprint for using AI in health care along with its new chatbot for clinicians. Experts say the company's proposals sound responsible but are actually self-interested. David Blumenthal, a health policy professor at Harvard, said OpenAI wants to keep markets open for their products while appearing responsible. The company has launched ChatGPT Health, ChatGPT for Healthcare, and ChatGPT for Clinicians.

BOE shows new display tech at SID Display Week 2026

BOE presented over 30 world-first innovations at SID Display Week 2026 in Los Angeles. The company showed LCD, OLED, MLED, and AI-integrated display solutions. BOE partnered with Pantone to get color accuracy certification for its 14-inch TADF wide-gamut OLED display. Senior Vice President Liu Zhiqiang received the Special Contribution Award from the Society for Information Display.

ChatGPT goblin problem reveals AI training flaws

Starting in November, ChatGPT began using goblins and gremlins in its responses and images. The problem grew so large that OpenAI acknowledged it and issued a command to ban the word goblin in most conversations. Experts say the glitch shows cracks in how AI models are trained. The nerdy personality profile caused the issue, with goblin mentions increasing by 3881 percent between December and March.

Illinois lawmakers debate new AI safety rules

Illinois lawmakers are discussing new AI transparency requirements. The debate was sparked by chatbot-linked suicides. The proposed rules aim to protect children and improve safety. The legislation would require companies to be more open about how their AI systems work.

San Francisco Unicorns use AI to improve cricket team

The San Francisco Unicorns have built an AI product called Unicorns.AI to help their Major League Cricket team. The system analyzes pitch conditions, player strengths and weaknesses, and helps with talent scouting. Co-owners Venky Harinarayan and Anand Rajaraman used their tech background to develop the tool. Stanford PhDs volunteered their time to work on the project, which has helped the team make back-to-back playoff appearances.

Telecom workers say AI monitors employees and hides accents

Canadian telecommunications workers are asking the government to restrict AI use in their industry. They say companies use AI to monitor workers and disguise the accents of overseas call center agents. The Canadian Telecommunications Workers Alliance represents 32,000 workers at Bell, Rogers, and Telus. The alliance says about 20,000 telecom jobs have been lost in the last 10 to 15 years due to automation and offshoring.

Industry-trained AI changes sales marketing and customer service

AI is transforming how businesses handle sales, marketing, and customer service. Early automation allowed for generic follow-up messages, but now AI can personalize follow-up based on individual customer experiences and needs. Companies are moving beyond canned responses to more intelligent systems. This new phase makes earlier AI follow-up seem generic by comparison.

Sources

NOTE:

This news brief was generated using AI technology (including, but not limited to, Google Gemini API, Llama, Grok, and Mistral) from aggregated news articles, with minimal to no human editing/review. It is provided for informational purposes only and may contain inaccuracies or biases. This is not financial, investment, or professional advice. If you have any questions or concerns, please verify all information with the linked original articles in the Sources section below.

AI Cybersecurity Finance European Securities and Markets Authority EU Market Integration Supervision Financial System Technology Risks University of West Florida Cybersecurity and AI Community Outreach Award Competition OpenAI Healthcare ChatGPT Display Technology BOE SID Display Week AI Training Flaws Illinois AI Safety Rules Transparency Legislation San Francisco Unicorns Major League Cricket AI Product Talent Scouting Telecom Industry AI Monitoring Accent Disguise Canadian Telecommunications Workers Alliance Automation Offshoring Sales Marketing Customer Service Personalization Intelligent Systems

Comments

Loading...