Artificial intelligence: new software category, education uses, health risks and financial fraud concerns

Spanish AI lab AiKit unveiled a new technology category called "Software for AIs" at a private summit in Alcalá de Henares, citing over 20,000 clients, a €5 million post‑tax profit in 2025 and a target of €10 million in 2026. The company describes the software as native‑AI‑operated business‑process management tools.

A commentary on generative AI noted that current models can recite facts but lack genuine humor, often reproducing existing jokes without creativity. Researchers highlighted statistical limits that prevent AI from improvising comedic content.

Education experts outlined six validated ways to harness generative AI for studying, including simplified explanations, auto‑generated quizzes, flashcard creation, real‑world examples, and rigorous fact‑checking to avoid hallucinations.

An international study of medical‑chat AI models (Gemini, DeepSeek, Meta AI, ChatGPT, Grok) found that about 50 % of answers to evidence‑based health queries were problematic, with 20 % highly risky. Grok performed worst (58 % highly problematic), while Gemini showed the fewest critical errors. Researchers warned that confident but inaccurate answers can mislead users and that AI‑generated references are often fabricated.

A comparative classroom experiment in Spain and Argentina used ChatGPT for assignments, measuring engagement, satisfaction, knowledge gain and transfer according to the Kirkpatrick model across 120 higher‑education students.

Mexico’s central bank, Banxico, warned that AI‑generated deepfakes enable large‑scale financial fraud, citing a case where a Hong‑Kong employee transferred $25 million after a fake video call. The bank cited a World Economic Forum estimate of a 1,740 % rise in deep‑fake fraud in North America and Deloitte’s projection of up to $40 billion in AI‑enabled fraud losses in the United States by 2027.

Research in basic education explored how generative AI can develop critical thinking, finding that AI tools can foster analytical skills but also risk passive reliance, recommending group activities that require students to explain AI‑generated ideas.