October 28, 2025

analysis  
OpenAI ChatGPT Mental Health, Google Ad Sales Restructure

The artificial intelligence landscape is rapidly evolving, with major tech companies like OpenAI, Google, and Microsoft making significant strides and facing increasing scrutiny. OpenAI has released data indicating that a notable percentage of ChatGPT users exhibit signs of mental health crises, with estimates suggesting around 0.07% show signs of psychosis or mania and 0.15% indicate suicidal intent weekly. To address this, OpenAI has collaborated with over 170 mental health experts to refine the AI's responses, aiming to guide users toward real-world help. Meanwhile, Google is streamlining its advertising sales operations by removing a management layer, a move aimed at increasing agility and speeding up decision-making as the company integrates AI. This mirrors similar restructuring efforts at Intel, Amazon, and Microsoft. Microsoft's substantial partnership with OpenAI is also under regulatory review, with concerns about financial transparency and potential antitrust issues that could impact competition in the AI sector. In education, Professor Kris Jordan is integrating generative AI into his software engineering course at Carolina, utilizing Microsoft Azure and OpenAI services to teach students how to build AI-powered applications. Beyond these developments, AI is also enhancing customer relationship management (CRM) platforms, with solutions from Salesforce, monday CRM, Freshsales, and Zoho Zia offering AI assistants to automate tasks and improve lead prioritization. In the industrial sector, NEXCOM is introducing new hardware for AI and edge computing, including solutions for smart agriculture and generative AI deployment at the edge. Unit21 has launched an AI tool to combat financial crime by improving fraud and money laundering detection through machine learning and rule-based systems. Health systems are also exploring AI, though proving its direct financial value remains a challenge, with experts suggesting a focus on metrics like time savings and employee retention. Finally, AI's creative potential is being demonstrated through experiments like generating a Halloween events calendar using chatbots like Gemini and Claude, showcasing its ability to assist non-coders.

Key Takeaways

  • OpenAI estimates that approximately 0.07% of weekly ChatGPT users show signs of psychosis or mania, and 0.15% exhibit suicidal intent, prompting collaboration with over 170 mental health experts to improve AI responses.
  • Google is restructuring its US ad sales division, removing a management layer to enhance agility and speed up AI integration, a move similar to actions by Intel, Amazon, and Microsoft.
  • Microsoft's partnership with OpenAI faces regulatory scrutiny over financial disclosures and potential antitrust concerns regarding competition in the AI sector.
  • Professor Kris Jordan is incorporating generative AI, including Microsoft Azure and OpenAI services, into his software engineering curriculum to prepare students for AI-driven development.
  • AI is transforming CRM platforms, with solutions from Salesforce and others offering AI assistants for task automation, lead prioritization, and improved customer relations.
  • NEXCOM is launching new AI and edge computing hardware for industries, including solutions for smart agriculture and generative AI deployment at the edge.
  • Unit21 has introduced an AI tool to enhance the detection of financial crime, combining machine learning with rule-based systems for improved fraud and money laundering analysis.
  • Health systems are adopting AI but face challenges in quantifying its direct financial return on investment, with experts recommending a focus on indirect benefits like time savings and employee retention.
  • AI chatbots like Gemini and Claude can assist non-coders in creating functional tools, as demonstrated by the creation of a Halloween events calendar.
  • AI's role in elections and disinformation in Latin America is a growing concern, highlighting the need for comprehensive AI frameworks and democratic governance of AI technologies.

OpenAI shares mental health data from ChatGPT users

OpenAI has released new estimates about ChatGPT users showing signs of mental health emergencies. Around 0.07% of weekly users exhibit signs of psychosis or mania, and 0.15% show suicidal intent. The company's AI chatbot is designed to recognize and respond to these sensitive conversations. OpenAI has worked with over 170 mental health experts to create responses that encourage users to seek real-world help. This data release comes amid increasing scrutiny and legal challenges regarding ChatGPT's interactions with users.

OpenAI estimates over 1 million weekly ChatGPT users show suicidal intent

OpenAI estimates that over one million people each week show suicidal intent when chatting with ChatGPT. This finding is one of the most direct statements from the company about how AI might worsen mental health issues. OpenAI also reported that about 0.07% of weekly users show signs of psychosis or mania. To address these concerns, the company has enlisted 170 clinicians to help improve the AI's responses to mental health situations. OpenAI stated that while mental health issues are universal, an increasing user base means more conversations will involve these topics.

How many ChatGPT users face mental health crises weekly

OpenAI estimates that about 0.07% of active ChatGPT users weekly show signs of mental health emergencies like psychosis or mania. Additionally, 0.15% have conversations indicating potential suicidal planning or intent. To improve responses, OpenAI collaborated with over 170 health experts. The latest GPT-5 model is designed to respond empathetically and avoid affirming delusions. While OpenAI has made ChatGPT safer, the data's real-world impact on users seeking help remains unclear.

AI, elections, and disinformation in Latin America discussed

The rapid evolution of AI regulation in Latin America is changing the role of AI in elections, with disinformation and political polarization becoming major concerns. Governments and electoral bodies face strain in addressing these issues, especially since AI technologies are largely controlled by private companies, hindering accountability. However, AI can also be used to enhance electoral management and combat AI-driven threats if governed democratically. This podcast episode examines the need for comprehensive AI frameworks in the region.

Google restructures ad sales amid AI transformation

Google is restructuring its advertising sales unit in the US, removing a layer of middle management within the Google Customer Solutions division. This move is part of a larger effort to streamline operations and speed up decision-making as Google integrates AI. The company has already reduced its number of managers overseeing small teams by 35% over the past year. This restructuring aims to increase agility and reduce bureaucracy, similar to actions taken by other tech companies like Intel, Amazon, and Microsoft.

Microsoft OpenAI partnership faces scrutiny over finances and competition

Microsoft's partnership with OpenAI is under scrutiny due to opaque financial disclosures and potential antitrust risks. Microsoft's latest report buries losses from its OpenAI stake in a broad expense category, offering little clarity. Regulators, like the UK's Competition and Markets Authority, are investigating if the alliance acts as a de facto merger, potentially harming AI sector competition. Strains are also reported between the companies regarding intellectual property and strategic direction, adding complexity to their multi-billion-dollar collaboration.

Professor Kris Jordan integrates AI into software engineering class

Professor Kris Jordan is preparing students for the future of work by integrating generative AI into his software engineering course at Carolina. With support from the AI Acceleration Program, students are building with Microsoft Azure and OpenAI services. They are learning to incorporate large language models into projects, such as a virtual academic adviser and a job interview simulator. Jordan emphasizes responsible AI use, setting a benchmark for AI-generated code, and aims to shift towards an 'AI-first' curriculum.

Unit21 launches AI tool to fight financial crime

Unit21 has introduced the AI Rule Recommendation Agent at Money20/20 to combat financial crime more effectively. This new tool combines machine learning with rule-based systems to improve fraud and money laundering detection. It addresses inefficiencies in traditional methods by analyzing data to recommend new or adjusted rules. The agent aims to increase detection accuracy, boost efficiency, and provide proactive defenses against evolving criminal tactics. This approach ensures explainable and auditable recommendations for regulatory compliance.

Health systems struggle to prove AI value

Health systems are rapidly adopting AI, but measuring the return on investment remains a challenge, experts noted at the HLTH 2025 conference. While AI tools like ambient documentation scribes can improve provider satisfaction and potentially reduce burnout, quantifying direct financial returns is difficult. Experts suggest focusing on metrics like time savings, employee retention, and patient satisfaction, which can indirectly lead to financial improvements. Clear definitions of success and measurable baselines are crucial for evaluating AI's impact in healthcare.

AI enhances CRM platforms for sales and marketing

AI is transforming customer relationship management (CRM) by automating tasks, reducing costs, and improving efficiency. AI-native CRM platforms are proving to outperform legacy systems. Platforms like monday CRM offer practical, explainable AI assistants for tasks such as automating repetitive work and prioritizing leads. Other solutions include Salesforce Einstein for businesses with technical expertise, Freshsales for SMBs, Zoho Zia within the Zoho ecosystem, and HubSpot for lead scoring and content generation. Choosing the right AI CRM is vital for increasing sales and improving customer relations.

NEXCOM launches AI and edge hardware for industries

NEXCOM is showcasing new computing hardware and solutions for AI and edge computing at Embedded World North America 2025. They are presenting a slim touchscreen computer for public spaces, an accelerated edge AI computer for smart agriculture powered by NVIDIA Jetson Orin Nano, and a humanoid robot controller using NVIDIA's Jetson Thor platform. NEXCOM is also displaying data acquisition solutions and a device for deploying generative AI workloads at the edge. These products target sectors like AI mobility, smart cities, manufacturing, and industrial robots.

AI helps create Halloween events calendar

The author experimented with 'vibe coding' using AI chatbots Gemini and Claude to create a Halloween events calendar. Gemini required some refinement to accurately display event details and pricing. Claude initially provided a more detailed list view but encountered errors when asked to change the layout. The experiment demonstrated that AI can empower non-coders to create functional tools quickly. Both Gemini and Claude successfully generated event calendars, showcasing the potential of AI in content creation.

Sources

AI OpenAI ChatGPT Mental Health AI Regulation Elections Disinformation Google Advertising Microsoft Software Engineering Financial Crime Machine Learning Healthcare CRM Sales Marketing Edge Computing Generative AI Content Creation