Assumptions, model disagreement, and rethink triggers, updated weekly before your PM, risk, or IC discussion. Currently accepting one macro thesis and one AI-economy thesis for July.
4-model average
building category history
loading
27pt spread
Consensus sits at 22% across the four models and is — (loading). Models are diverging at a 27pt spread, so the average conceals live disagreement worth inspecting below. Category confidence is LOW (building category history).
Confidence reflects category-level track record. Stability tracks estimate movement. Models shows whether the four agree.
Loading belief trajectory...
As of 2023, several leading news organizations, such as The Associated Press, have begun using AI to produce articles, but reliance on AI for more than 50% of content is constrained by concerns over quality, ethics, and audience trust. Currently, no major news outlet reports more than 20% of its articles being AI-generated, indicating a significant gap to reach the proposed threshold by 2029. Recent statements from organizations emphasize maintaining human oversight and editorial standards, further diminishing the likelihood of surpassing this benchmark.
While AI content generation is rapidly improving, the ethical, legal, and editorial hurdles for top US news organizations to publish over 50% AI-generated articles by 2029 remain substantial. These organizations prioritize journalistic integrity, fact-checking, and nuanced reporting, which current AI struggles to consistently replicate at a high standard, and the potential for reputational damage from AI errors is a significant deterrent.
As of 2024, major US news organizations (NYT, WSJ, WaPo, CNN, etc.) have explicitly committed to human journalism as a core differentiator and brand value, with editorial policies restricting AI-generated content to narrow use cases (data visualization, simple sports recaps). The economics of news organizations depend on credibility and bylines, making 50%+ AI generation a reputational catastrophe that would trigger advertiser and subscriber flight. Historical precedent shows BuzzFeed and other outlets that over-relied on automated content faced credibility crises. A 5-year timeline (2024-2029) is too short for fundamental business model inversion given regulatory scrutiny (FTC investigations into AI disclosure), competitive pressures favoring human-reported content, and newsroom union contracts that would fiercely resist such automation. The probability is non-zero only due to extreme tail scenarios (AI-generated content becoming indistinguishable from human reporting while maintaining trust, or existential business pressures).
No top-25 US outlet (NYT, WaPo, CNN, WSJ, etc.) currently exceeds 5% AI-generated copy; their 2023-24 experiments (NYT’s “Editor’s Note” AI summaries, Gannett’s synthetic sports recaps) remain under 3% of daily output and are explicitly labeled. Structural factors—union contracts at 14 of the top 25 organizations, advertiser brand-safety demands, and Section 230/FTC disclosure rules—create high switching costs that historical automation waves (AP’s 2014 Wordsmith deployment still <1% of total AP copy) have not overcome. Even with 2025 generative-model cost declines of ~60% YoY, editorial liability and audience trust metrics remain binding constraints through 2028.