Last update: Tuesday 10/7/25
ChatGPT's TL;DR summaries of Top 3 stories ...
1. Sora | 2. California | 3. PJM
1) "OpenAI’s Sora Makes Disinformation Extremely Easy and Extremely Real", Tiffany Hsu, Stuart A. Thompson, and Steven Lee Myers, NY Times, 10/3/24 ***
-- This story also covered by TechCrunch, WSJ, Gizmodo, Engadget, NPR, ... and OpenAI- Text Gizmodo ... Contains many sample videos
*** 1, The First 24 Hours of Sora 2 Chaos: Copyright Violations, Sam Altman Shoplifting, and More
a) Widespread copyright chaos on launch
Within a day of Sora 2’s release, users flooded social media with videos featuring copyrighted characters such as Sonic, Pikachu, and Rick and Morty. OpenAI’s “opt-out” policy allows copyrighted material to appear in generated videos unless rights holders explicitly prohibit it.
Within a day of Sora 2’s release, users flooded social media with videos featuring copyrighted characters such as Sonic, Pikachu, and Rick and Morty. OpenAI’s “opt-out” policy allows copyrighted material to appear in generated videos unless rights holders explicitly prohibit it.
-
Disney has already opted out, blocking its characters from use.
-
Evidence suggests the model was trained on copyrighted material despite lacking permission.
b) Model behavior reveals embedded bias toward known IP
Even generic prompts produce outputs resembling recognizable works like The NeverEnding Story, showing how deeply Sora 2 internalized copyrighted imagery and styles. The model seems unable to avoid reproducing familiar creative assets.
Even generic prompts produce outputs resembling recognizable works like The NeverEnding Story, showing how deeply Sora 2 internalized copyrighted imagery and styles. The model seems unable to avoid reproducing familiar creative assets.
-
Indicates significant overlap between training data and copyrighted media.
-
Raises questions about OpenAI’s ability—or willingness—to filter outputs effectively.
c) The Sam Altman “cinematic universe”
OpenAI allows verified users to upload their likeness for video generation, and CEO Sam Altman’s own image has become a meme—appearing in clips of him shoplifting GPUs, turning into a cat, and parodying theft from famed directors.
-
Altman’s likeness is available for all users; others can set tighter controls.
-
Highlights both engagement potential and reputational risks for public figures.
d) Early safeguards fall short
While OpenAI embeds metadata and watermarks to mark videos as AI-generated, users can still crop or export content, defeating these protections. Once a video leaves Sora, OpenAI loses control over its distribution.
While OpenAI embeds metadata and watermarks to mark videos as AI-generated, users can still crop or export content, defeating these protections. Once a video leaves Sora, OpenAI loses control over its distribution.
-
Users can revoke in-app access, but copies persist outside the platform.
-
Points to looming legal and social fallout over likeness misuse and copyright violations.
2) "California Governor Signs Sweeping A.I. Law", Cecilia Kang, NY Times, 9/29/25 ***
-- This story also covered by Politico, Reuters, NBC News (video) ... and Governor Newsom- Text NY Times
*** 2. California Governor Signs Sweeping A.I. Law
a) California enacts one of the nation’s strongest A.I. safety laws
a) California enacts one of the nation’s strongest A.I. safety laws
Governor Gavin Newsom signed the Transparency in Frontier Artificial Intelligence Act (S.B. 53), establishing rigorous safety and transparency requirements for developers of advanced A.I. systems. The law positions California at the forefront of A.I. regulation in the U.S.
-
Companies must report their safety protocols and disclose major risks associated with their models.
-
The law strengthens whistle-blower protections for employees exposing potential dangers.
b) Focus on high-revenue A.I. firms and frontier models
The law applies to companies generating at least $500 million annually and developing cutting-edge A.I. technologies. It mandates public disclosure of how these companies incorporate best safety practices aligned with international standards.
The law applies to companies generating at least $500 million annually and developing cutting-edge A.I. technologies. It mandates public disclosure of how these companies incorporate best safety practices aligned with international standards.
-
Firms must report any safety incidents to California’s Office of Emergency Services.
-
The goal is to make high-impact A.I. systems more accountable and auditable.
c) Tech industry pushback and calls for federal preemption
Major tech players—Meta, OpenAI, Google, and Andreessen Horowitz—oppose state-level A.I. laws, warning that a patchwork of regulations could stifle innovation. They are lobbying for a single federal framework that would override state initiatives.
Major tech players—Meta, OpenAI, Google, and Andreessen Horowitz—oppose state-level A.I. laws, warning that a patchwork of regulations could stifle innovation. They are lobbying for a single federal framework that would override state initiatives.
-
Meta and Andreessen Horowitz pledged $200M to super PACs supporting A.I.-friendly candidates.
-
Industry groups argue the law risks creating inconsistent national standards.
d) A scaled-back version of last year’s vetoed bill
Newsom previously vetoed a more aggressive bill that included mandatory safety testing and a “kill switch” for dangerous A.I. systems. The new version, rewritten with input from academics and policy experts, is viewed as a more pragmatic compromise.
Newsom previously vetoed a more aggressive bill that included mandatory safety testing and a “kill switch” for dangerous A.I. systems. The new version, rewritten with input from academics and policy experts, is viewed as a more pragmatic compromise.
-
The revised law focuses on transparency and accountability rather than direct technical controls.
-
Lawmakers describe it as a “reasonable approach” balancing innovation and risk.
e) Continued leadership by California in tech regulation
California once again leads where Congress has stalled, following earlier moves on data privacy and children’s online safety. The state will also form an A.I. consortium under its Government Operations Agency to promote ethical and sustainable A.I. research.
California once again leads where Congress has stalled, following earlier moves on data privacy and children’s online safety. The state will also form an A.I. consortium under its Government Operations Agency to promote ethical and sustainable A.I. research.
-
Reflects California’s role as both an innovation hub and regulatory trendsetter.
-
Other states may follow—38 have already introduced A.I. bills this year.
f) Support from Anthropic and A.I. safety advocates
Unlike many peers, Anthropic backed the legislation, calling it a “practical safeguard” that enforces accountability while supporting responsible innovation.
Unlike many peers, Anthropic backed the legislation, calling it a “practical safeguard” that enforces accountability while supporting responsible innovation.
-
Jack Clark of Anthropic said the law will help ensure safer A.I. development as capabilities accelerate.
-
Signals a split in the A.I. industry between safety-first firms and those resisting oversight.
3) "Data center boom sparks sticker shock for PJM ratepayers", Peter Behr, E&E News/Politico, 10/3/25 ***
- Text E&E News/Politico
*** 3. Data center boom sparks sticker shock for PJM ratepayers
a) Data center expansion is driving massive electricity demand across PJM
The PJM Interconnection—covering 13 states from Illinois to New Jersey—faces an unprecedented surge in electricity demand from data centers, adding nearly 12,000 megawatts to projected peak demand next summer. Most of this increase comes from large-scale AI and cloud computing operations.
-
PJM’s July capacity auction jumped 82%, adding $7.2 billion in future power payments.
-
These costs are passed directly to ratepayers across the region.
b) Billions in transmission costs are being shifted to consumers
According to the Union of Concerned Scientists (UCS), utilities launched over 150 local transmission projects between 2022 and 2024—almost entirely to connect new data centers.
According to the Union of Concerned Scientists (UCS), utilities launched over 150 local transmission projects between 2022 and 2024—almost entirely to connect new data centers.
-
Consumers absorbed $4.3 billion in transmission expansion costs across seven PJM states last year.
-
The UCS criticized weak regulatory oversight that allows utilities to recover costs with “practically nonexistent” review.
c) Rising electricity prices and long-term inflation risk
The Natural Resources Defense Council (NRDC) projects PJM ratepayers could face $163 billion in higher electricity costs through 2033.
The Natural Resources Defense Council (NRDC) projects PJM ratepayers could face $163 billion in higher electricity costs through 2033.
-
The average household may pay $70 more per month by 2028 due to capacity cost increases.
-
These costs persist even if some data centers are delayed or canceled, since capacity payments are locked in.
d) The policy debate: who should generate the power?
Experts including NRDC and PJM’s market monitor, Joseph Bowring, argue data center developers should be required to provide their own power generation and storage rather than relying solely on the grid.
Experts including NRDC and PJM’s market monitor, Joseph Bowring, argue data center developers should be required to provide their own power generation and storage rather than relying solely on the grid.
-
PJM initially considered a mandatory “bring-your-own-generation” rule but retreated after industry pushback.
-
A voluntary version is now under discussion, with final proposals expected by November 19.
e) PJM’s strained interconnection process
PJM has a backlog of interconnection requests totaling 46,000 MW, forcing a temporary shutdown of its approval process. Although many projects have now cleared technical review, they remain delayed by financing and permitting barriers.
PJM has a backlog of interconnection requests totaling 46,000 MW, forcing a temporary shutdown of its approval process. Although many projects have now cleared technical review, they remain delayed by financing and permitting barriers.
-
PJM insists it lacks authority to veto new customers like data centers seeking grid access.
-
Its staff is exploring ways to track “double-counting” by developers proposing identical projects in multiple regions.
f) Broader implications: inflation, grid strain, and governance challenges
The surge in AI-driven data center demand threatens to keep power prices elevated for a decade, potentially fueling persistent inflation and straining grid reliability.
The surge in AI-driven data center demand threatens to keep power prices elevated for a decade, potentially fueling persistent inflation and straining grid reliability.
-
Outgoing PJM CEO Manu Asthana framed the issue as part of the U.S. “race for AI dominance.”
-
Analysts warn that without coordinated national planning, the AI boom could leave consumers paying for speculative infrastructure they may never benefit from.
No comments:
Post a Comment
Your comments will be greatly appreciated ... Or just click the "Like" button above the comments section if you enjoyed this blog note.