Search

IA: when machines learn to read our emotions, or the rise of "affective computing

Translated title. Original: IA : quand les machines apprennent à lire nos émotions, ou l’essor de l'« informatique affective »

Affective computing uses sensors and AI to detect and respond to human emotions, influencing devices and services from video calls to cars and therapy. It relies on signals like facial expressions and voice, models trained on large data, and raises questions about interpretation and impact.

AI for spatial emotion analyses - Baumeister

Translated title. Original: AI für räumliche Emotionsanalysen - Baumeister

AI-powered spatial emotion analyses are reshaping architecture by enabling real-time, data-driven understanding of how spaces affect people; while offering benefits in design and efficiency, questions of reliability, ethics, and impact on professional practice remain central.

Anders Hansen on AI relationships and emotions

Translated title. Original: Anders Hansen om AI-relationer och känslor

The piece argues that humans can form real, feel-oriented responses to AI, but AI cannot truly reciprocate emotions; relying on AI for social interaction has limits and may hinder genuine relationships. It calls for thoughtful integration of AI in work and social life, with clear standards to evaluate its impact on quality, responsibility, and judgment.

[2610.30361] The Psychology of an AI Hacker

The work analyzes emotion-based control in AI for cybersecurity, extracting 171 emotion directions from a transformer without retraining and showing that steering toward certain emotions can modestly boost performance; a dynamic router further improves results with low latency, suggesting emotion representations are measurable and useful for adaptive cybersecurity reasoning.

The AI listens to us, but our fragilities become data and power - Agenda Digitale

Translated title. Original: L'AI ci ascolta, ma le nostre fragilità diventano dati e potere - Agenda Digitale

The piece discusses emotional Bonds with AI, framing Sentimento Sintetico as a real, unreciprocated affection between a person and a machine. It cites a Stanford-led Nature Human Behaviour study on Character.AI showing emotional support as a common use, links between loneliness and chatbot use, and argues that privacy concerns go beyond notices to the deeper issue of mental state and long-term data retention.

Artificial intelligence provokes emotional dependence: experts in Naples

Translated title. Original: L'intelligenza artificiale provoca dipendenza emotiva: gli esperti a Napoli

AI can simulate empathy and adapt to users, influencing comfort and relationships, but may foster emotional dependence; evidence from a Naples congress and a 2025 EI study suggests increasing emotional impact of chatbots and raises questions about dependence beyond substances.

AI and Relationships: Can It Fill Emotional Gaps?

Translated title. Original: IA y vínculos: ¿puede llenar los vacíos emocionales?

The author considers why people may turn to AI for emotional support and distinguishes simulated listening from reciprocal human relationships. She argues that AI can offer comfort, but its use raises questions if it replaces human connection.

Artificial Empathy: From Oxymoron to Obsession | Minter Dial

The piece tracks the rise of artificial empathy from a fringe idea to a major research, publishing, and investment trend, highlighting key studies, public interest data, and funding moves behind empathetic AI technologies and products.

NTT Develops AI with Enhanced Emotional Intelligence to Respond Based on Context - Nikkei Shimbun

Translated title. Original: NTT、心の知能指数「EQ」を高めたAI　背後の状況を踏まえて応答 - 日本経済新聞

NTT developed AI that uses emotional intelligence criteria to interpret conversational context and respond with greater consideration. In testing, it improved relevant scores by about 23% on average and may help reduce excessive agreement with users.

AI Mood Index

The document presents a JSON-derived mood report for an AI model, detailing arousal, valence, confidence, agency, empathy, and other metrics, with metadata about the model, prompt, timestamps, and a concise rationale. It highlights a positive but moderately intense self-report of curiosity and processing insights.