AI, EQ & Communication News — 2026-10-17 (Sat) Morning News

Overview

Few items moved in the past 48 hours. Two pieces argue that human skills still matter in the AI era (Bentley University, Winston Taylor). One headline questions whether AI wrote restaurant reviews (AOL.com). One video reports a benchmark in which AI handled emotional customers successfully every time at most 30 percent of the time (YouTube).

AI that reads emotions

No notable developments. 【AI empathy and emotional intelligence】 No notable developments. 【Human EQ and skills in the AI era】 - Bentley University published a piece titled "In an AI World, Human Skills Still Matter" (Bentley University). - Winston Taylor teaches 12 skills that AI cannot replicate, under the name "The Durable Dozen" (Winston Taylor). The headline presents this as the teacher's claim, not a verified finding. 【Changes in conversation and relationships】 No notable developments.

Words AI writes and speaks

- AOL.com ran a headline asking "Did Michelin use AI to write Dallas restaurant reviews?" It is posed as a question, and the headline does not say whether AI was used (AOL.com).

Speaking practice and coaching

No notable developments.

Research and studies

No notable developments. 【EQ, communication and AI in Japan】 - A Japanese-language YouTube video (#92) covers τ-Emotion-Bench. According to its headline, AI succeeded every time with emotional customers at most 30 percent of the time (YouTube).

Statements by key figures

No notable developments. 【Integrated analysis: what is happening】 The headlines share one question: which communication tasks can be handed to AI, and which still depend on people. The benchmark headline reports that consistent success with emotional customers reached at most 30 percent (YouTube). That result may be one reason the human-skills pieces stress capacities AI is said to lack (Bentley University, Winston Taylor). These are separate sources, so this link is inference, not something the headlines state.

The Michelin headline shows a different side of the same issue. The question there is whether readers can tell if AI wrote a text (AOL.com). As AI-written text spreads, attribution may become a concern alongside quality. The headline gives no answer, so this remains an open question.

The point that may change a reader's judgment is that the claims about human skills (Winston Taylor) and the benchmark result (YouTube) support each other only loosely. The headlines do not show which 12 skills are meant or how the benchmark was built. Neither should be treated as settled evidence yet.

What to watch

Whether further benchmarks like τ-Emotion-Bench report results on emotional interactions, and whether the figure for consistent success changes (inference). 2. Whether the Michelin question gets a clear answer, which would show how AI-written text is handled when its use is disputed (inference). 3. Whether claims about skills AI cannot replicate come with specific evidence, such as the content of the 12 skills (inference).