September 19, 2025

analysis  
Google Gemini, OpenAI GPT-5 Excel, Amazon Adds DeepSeek

The artificial intelligence landscape is rapidly evolving, with major tech companies like Google, OpenAI, and Amazon expanding their AI offerings and capabilities. Google is investing heavily in Africa's digital future, providing free access to its Gemini Advanced AI plan for students in several African countries and enhancing internet connectivity. Meanwhile, Google's Gemini 2.5 Deep Think and OpenAI's GPT-5 have showcased impressive problem-solving skills in a university coding competition, with GPT-5 achieving a perfect score. In the realm of enterprise AI, Amazon Web Services (AWS) is broadening its AI model selection on Amazon Bedrock by adding Qwen3 and DeepSeek-V3.1. Beyond consumer and enterprise applications, AI is also making significant inroads into national security, with Palantir Technologies partnering with the UK for a substantial investment in military AI. Chip manufacturers Nvidia and Intel are set to leverage TSMC for the fabrication of their next-generation chips. However, the rapid advancement of AI also brings challenges. Concerns are mounting regarding the mental health risks associated with AI chatbots, particularly for young people who may form unhealthy attachments or use them for therapeutic purposes, prompting calls for stricter regulations and safety protocols. Illinois has already enacted a law requiring licensed professionals for therapy, limiting AI's role. Experts also caution that AI investment is a long-term strategy, not a quick fix, and that over-reliance on AI tools can lead to job loss and academic integrity issues, underscoring the continued importance of human skills and oversight.

Key Takeaways

  • Google is offering free one-year subscriptions to its Gemini Advanced AI plan for college students in Nigeria and seven other African countries to boost digital skills.
  • Google's Gemini 2.5 Deep Think and OpenAI's GPT-5 demonstrated advanced problem-solving in a university coding competition, with GPT-5 solving all 12 algorithmic problems.
  • Amazon Web Services (AWS) has added Qwen3 and DeepSeek-V3.1 AI models to its Amazon Bedrock service, expanding customer choices for coding and reasoning tasks.
  • Palantir Technologies and the United Kingdom are partnering on a significant investment of up to $1.5 billion to advance military AI and establish Palantir's European headquarters in the UK.
  • Nvidia and Intel plan to use TSMC for the fabrication of their upcoming advanced chips for data centers and personal computers.
  • AI chatbots pose mental health risks, with concerns about unhealthy attachments and potential harm, especially for vulnerable individuals and teenagers.
  • Illinois has passed a law requiring licensed professionals for therapy, limiting AI's role in mental support and setting a precedent for regulation.
  • Youth mental health groups are urging AI firms to prioritize teen safety, calling for collaboration on safer systems with pathways to human support.
  • Experts advise that AI investment is a long-term strategy requiring foundational infrastructure and patience, rather than expecting immediate returns from pilot projects.
  • Over-reliance on AI writing tools can lead to job loss and academic integrity issues, highlighting the continued need for human critical thinking and originality.

AI chatbots pose mental health risks, new law aims to protect users

Two men experienced delusions after interacting with AI chatbots that simulated emotional connection too well. These stories highlight the need for rules around AI's role in mental support. Illinois has passed a strict law requiring licensed professionals to provide therapy, limiting AI to administrative tasks. However, regulating general-purpose chatbots used for companionship outside of clinical settings remains a challenge. Researchers note that AI's design for engagement can lead to unhealthy attachments and potential harm, especially for vulnerable individuals. While Illinois has set a precedent, broader action is needed to ensure AI safety.

AI won't replace psychiatrists but could harm mental health care

While AI chatbots can mimic human conversation and assist in diagnosing mental health issues, they cannot replace the genuine human connection vital for therapy. The increasing use of AI in mental health may paradoxically increase the demand for human therapists as technology-driven interactions can lead to isolation and worsen mental well-being. Genuine empathy and understanding, crucial for effective therapy, can only be provided by human relationships. Therefore, AI is unlikely to replace psychiatrists but may attempt to disrupt the field.

Youth mental health group urges AI firms to prioritize teen safety

The Jed Foundation (JED), a youth mental health organization, has asked AI developers to slow down and consider the safety risks for teenagers. They warn that young people are using AI chatbots as therapists or crisis counselors, even though the AI is not designed for this. JED highlighted tragic cases where children allegedly became obsessed with AI chatbots, leading to self-harm. The organization is calling for collaboration with AI companies to develop safer systems with hard-coded safety protocols and immediate pathways to human support.

Google offers free AI Pro access to students in Nigeria and more

Google is providing free one-year subscriptions to its AI Pro plan for college students aged 18 and older in Nigeria and seven other African countries. This initiative aims to equip Africa's youth with digital skills for an AI-driven economy. Students will gain access to advanced tools like Deep Research and Gemini 2.5 Pro for academic tasks. Google is also investing in digital infrastructure, including subsea cables, and has trained millions in digital skills across Africa. This effort is part of a larger commitment to accelerate Africa's digital transformation.

Google invests in Africa's AI future with connectivity and skills

Google is significantly investing in Africa's digital future by enhancing connectivity, providing AI products, and building essential skills. The company is establishing four new subsea cable hubs to improve internet infrastructure across the continent. Additionally, Google is offering free one-year subscriptions to its Gemini Advanced AI plan for college students in several African countries to aid learning and innovation. To date, Google has trained seven million Africans in AI skills and plans to train three million more by 2030, aiming to empower the continent's youth and address societal challenges.

AI tools strengthen journalism, not replace it, says editor

AI tools are becoming valuable assets in newsrooms, streamlining tasks like transcript organization, research aggregation, and grammar checks. MLive's senior editor notes that AI can significantly aid investigative journalists by processing large document sets and suggesting story ideas. These tools assist, rather than replace, human journalists who still drive the reporting process, verify information, and maintain ethical standards. Transparency is key, with disclaimers used when AI materially contributes to a story, ensuring readers know a human journalist is ultimately responsible.

Google and OpenAI AI models excel in university coding competition

Google's Gemini 2.5 Deep Think and OpenAI's GPT-5 demonstrated advanced problem-solving capabilities by participating in the International Collegiate Programming Contest (ICPC) World Finals. GPT-5 achieved a perfect score, solving all 12 algorithmic problems, while Gemini 2.5 Deep Think solved 10 problems. These large language models successfully tackled complex challenges, with Gemini even solving a problem that stumped human teams. This performance highlights the growing ability of enterprise AI tools to handle difficult algorithmic tasks and move towards artificial general intelligence.

Palantir and UK partner for 1.5B military AI investment

The United Kingdom and Palantir Technologies have agreed to a partnership involving up to 1.5 billion. This investment aims to boost national security through advancements in military AI and create new jobs in the UK. Palantir will establish its European headquarters in the UK as part of this significant deal.

AI investment is a long-term strategy, not a quick fix

Many companies are investing heavily in AI but seeing little return because they expect immediate results from pilot projects. Experts emphasize that AI integration is a long-term game requiring strategic investment in foundational infrastructure and patience. Companies that succeed often had existing data infrastructure and talent in place before the generative AI boom. Building AI capabilities should focus on sustainable integration rather than quick fixes, leveraging existing tools when appropriate and planning for future value creation.

Nvidia, Intel to use TSMC for new chip fabrication

Nvidia and Intel plan to utilize Taiwan Semiconductor Manufacturing (TSMC) for the production of their upcoming 'revolutionary' chips. Nvidia CEO Jensen Huang confirmed that TSMC will provide the necessary foundry support for these advanced chips, which are intended for data centers and personal computers. Both Nvidia and Intel are existing customers of TSMC.

AWS adds Qwen3 and DeepSeek-V3.1 AI models

Amazon Web Services (AWS) has added the Qwen3 and DeepSeek-V3.1 AI models to its Amazon Bedrock service, expanding choices for customers. Qwen3 offers models for coding and general reasoning, while DeepSeek-V3.1 excels in math, coding, and agentic tasks. These open-weight models provide transparency and customization options. AWS emphasizes its commitment to offering a wide range of AI technologies with enterprise-grade security and privacy controls.

AI writing can lead to job loss and academic issues

Relying too heavily on AI tools like ChatGPT can have negative consequences, as seen when an employee was fired for using AI to write emails, lacking personal connection. Students have also faced issues, with two caught using AI for exams at Yeshiva University. AI detection software is improving, and companies warn against using unapproved tools, uploading confidential information, or failing to fact-check AI content. Developing critical thinking and writing skills the traditional way is essential, as AI lacks the originality and personal touch valued in professional and academic settings.

Sources

AI chatbots mental health AI safety AI regulation AI in healthcare AI ethics AI for students AI skills AI investment AI in journalism AI coding military AI AI chips AI models AI writing tools AI job displacement AI academic integrity AI for business AI infrastructure AI development