Generative Artificial Intelligence has been adopted rapidly by consumers, with the article noting 1.8 billion users and a 39.4% adoption rate among U.S. adults in a two-year span. Despite that growth, willingness to pay is very low: Menlo Ventures’ State of Consumer AI report is cited showing only about 3% of users subscribe to premium tiers. The piece highlights a disconnect between fast consumer uptake and weak monetization, and it repeats that OpenAI and Anthropic remain far from profitability. The article also states that the global annual revenue figure for AI apps from those users was not provided in the original reporting and is marked as Not stated.
On the enterprise side, the article describes mounting skepticism. It cites an MIT report finding that 95% of enterprise Artificial Intelligence projects fail to substantially improve efficiency or profits, attributing the problem less to model performance and more to difficulties integrating models with company data, workflows, and infrastructure. The coverage notes that many technology leaders are talking less about artificial general intelligence and superintelligence and more about ensembles of specialized models. The author points out that some tasks, such as coding, are being made more efficient, but few chief technology officers claim generative Artificial Intelligence is yet transforming their businesses in measurable ways.
The article also covers legal and industry shifts, reporting that Anthropic has agreed to preliminary settlement terms in a lawsuit brought by authors who alleged its Claude models were trained on copyrighted books obtained from shadow libraries. Judge William Alsup gave the parties until September 5 to finalize the settlement. Alsup’s earlier ruling found the use could be considered fair use while also saying the works had been obtained unlawfully from shadow libraries. The piece concludes that the honeymoon period for generative Artificial Intelligence may be ending as users and companies demand concrete productivity gains and workable integration rather than dazzlement alone.
