Deepfakes Erode Video Trust. Tado teaches thermostats to spot open windows. Claude Duet Acts Proactively. Few Companies See AI Value.

The AI news for October 23rd, 2026 Here are the details of the day's selected top stories: Study: AI videos massively undermine trust in moving images. Source: https://www.heise.de/news/Studie-KI-Videos-untergraben-Vertrauen-in-Bewegtbild-massiv-11497558.html?wt_mc=rss.red.ho.themen.k%C3%BCnstliche+intelligenz.beitrag.beitrag Tado: AI should better recognize open windows – heating pause still only in the subscription. Source: https://www.heise.de/news/Tado-KI-soll-offene-Fenster-besser-erkennen-Heizpause-weiter-nur-im-Abo-11495557.html?wt_mc=rss.red.ho.themen.k%C3%BCnstliche+intelligenz.beitrag.beitrag Anthropic is sending proactive long-running agents 'Claude Duet' into the race against Grok Bots, Meta Muse, and OpenAI's Dots. Source: https://the-decoder.de/anthropic-schickt-proaktiven-dauer-agenten-claude-duet-gegen-grok-bots-meta-muse-und-openais-dots-ins-rennen/ Gartner survey: Only 13 percent of companies derive real business value from AI investments. Source: https://the-decoder.de/gartner-umfrage-nur-13-prozent-der-unternehmen-ziehen-echten-geschaeftswert-aus-ki-investitionen/