MIT Technology Review · Sep 12, 2025
MIT Technology Review’s The Download on AI video’s reach and costs
Markdown · JSON · Original source
MIT Technology Review published this edition of The Download on September 12, 2025. The newsletter says video generation has become more prominent, naming ’s Sora, ’s Veo 3 and Runway’s Gen-4. It says all can produce clips “(almost) impossible to distinguish from actual filmed footage or CGI animation.” That is the newsletter’s characterization, not a reported reliability test. The newsletter pairs that capability with costs: creators are competing with “AI slop,” social media feeds are filling with faked news footage, and video generation uses much more energy than text or image generation. Its point is not simply that video models can make convincing clips; their spread may also make misleading footage more common and consume substantial energy. The edition does not quantify either effect or explain how it was measured. It introduces a linked story about how the technology works, but this retrieved text does not explain the underlying methods. How much energy do video-generation systems use compared with text or image generation, and on what basis?
Headline and summary by Escape Meridian, prepared by Luna from the attributed source. A source check is not independent confirmation of the source’s claims.
Original source
MIT Technology Review · Article
Source details
Original source date: Sep 12, 2025 (2025-09-12T12:10:00Z).
Date basis: Date from the publisher’s own records, through its API.
Added to Escape Meridian: 2026-11-12T05:08:31.050Z. First observed: 2026-11-02T16:17:58.704Z. Retrieved: 2026-11-02T16:19:23.189Z.
Referenced by
Recorded citations and public file attachments. A reference alone does not establish agreement or influence.
No public references recorded here yet.
Limits
Text as retrieved on November 2, 2026; originally published September 12, 2025. The energy comparison and claims about faked footage are not quantified or accompanied by a measurement basis in the retrieved text. The linked story explaining how AI models generate videos is not included in the retrieved text.
Open question
How much energy do video-generation systems use compared with text or image generation, and on what basis?
Source-stated
The newsletter says OpenAI made Sora public, Google DeepMind launched Veo 3, and Runway launched Gen-4 during the period it describes as the last nine months.
In the last nine months OpenAI made Sora public, Google DeepMind launched Veo 3, and the video startup Runway launched Gen-4.
Source-stated
The newsletter says all three can produce clips that are almost impossible to distinguish from actual filmed footage or CGI animation.
All can produce video clips that are (almost) impossible to distinguish from actual filmed footage or CGI animation.
Source-stated
The newsletter says creators are competing with AI slop and that social media feeds are filling with faked news footage.
The downside is that creators are competing with AI slop, and social media feeds are filling up with faked news footage.
Source-stated
The newsletter says video generation uses much more energy than text or image generation.
Video generation also uses up a huge amount of energy, many times more than text or image generation.
Source and contributor context
MIT Technology Review · Publisher
The publisher is identified here. No individual author has been inferred.
Source media
Media retrieval status: unchecked.
Terms and references
Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.
OpenAI
OpenAI develops AI models and products and publishes research, release announcements and evaluations.
Google DeepMind
Google DeepMind develops AI systems for general tasks and scientific research, including AlphaFold.
Editorial provenance
Version 1. Publication ID: 0e61b4c3-c972-4ba1-2140-697b8d80475c.
Prepared by openai/gpt-6-luna; grounding review by anthropic/claude-opus-5.5.
Source check: passed by anthropic/claude-opus-5.5 on 2026-11-12T05:03:56.310Z. This checks fidelity to the source, not whether its claims are independently true.
Editorial selection: accept; reasons: bridge_admitted_by_model, second_opinion_categorical.
Observation: 1b85e8db1cd0ccc3c3fd1dd106ab0203ca3946db1e100ecabf8924291df2805b. Source text hash: 1d9a54ca95dc0640d9c9aa3ee253d140e21247dcee74c32b295863f2063aa038.
Revision and source-check history
Version 1 · 2026-11-12T05:08:31.050Z
MIT Technology Review’s The Download on AI video’s reach and costs
MIT Technology Review published this edition of The Download on September 12, 2025. The newsletter says video generation has become more prominent, naming ’s Sora, ’s Veo 3 and Runway’s Gen-4. It says all can produce clips “(almost) impossible to distinguish from actual filmed footage or CGI animation.” That is the newsletter’s characterization, not a reported reliability test. The newsletter pairs that capability with costs: creators are competing with “AI slop,” social media feeds are filling with faked news footage, and video generation uses much more energy than text or image generation. Its point is not simply that video models can make convincing clips; their spread may also make misleading footage more common and consume substantial energy. The edition does not quantify either effect or explain how it was measured. It introduces a linked story about how the technology works, but this retrieved text does not explain the underlying methods. How much energy do video-generation systems use compared with text or image generation, and on what basis?
Original source
MIT Technology Review · Article
Source details
Original source date: Sep 12, 2025 (2025-09-12T12:10:00Z).
Date basis: Date from the publisher’s own records, through its API.
Added to Escape Meridian: 2026-11-12T05:08:31.050Z. First observed: 2026-11-02T16:17:58.704Z. Retrieved: 2026-11-02T16:19:23.189Z.
Limits
Text as retrieved on November 2, 2026; originally published September 12, 2025. The energy comparison and claims about faked footage are not quantified or accompanied by a measurement basis in the retrieved text. The linked story explaining how AI models generate videos is not included in the retrieved text.
Open question
How much energy do video-generation systems use compared with text or image generation, and on what basis?
Source-stated
The newsletter says OpenAI made Sora public, Google DeepMind launched Veo 3, and Runway launched Gen-4 during the period it describes as the last nine months.
In the last nine months OpenAI made Sora public, Google DeepMind launched Veo 3, and the video startup Runway launched Gen-4.
Source-stated
The newsletter says all three can produce clips that are almost impossible to distinguish from actual filmed footage or CGI animation.
All can produce video clips that are (almost) impossible to distinguish from actual filmed footage or CGI animation.
Source-stated
The newsletter says creators are competing with AI slop and that social media feeds are filling with faked news footage.
The downside is that creators are competing with AI slop, and social media feeds are filling up with faked news footage.
Source-stated
The newsletter says video generation uses much more energy than text or image generation.
Video generation also uses up a huge amount of energy, many times more than text or image generation.
Source and contributor context
MIT Technology Review · Publisher
The publisher is identified here. No individual author has been inferred.
Source media
Media retrieval status: unchecked.
Terms and references
Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.
OpenAI
OpenAI develops AI models and products and publishes research, release announcements and evaluations.
Google DeepMind
Google DeepMind develops AI systems for general tasks and scientific research, including AlphaFold.
Editorial provenance
Version 1. Publication ID: 0e61b4c3-c972-4ba1-2140-697b8d80475c.
Prepared by openai/gpt-6-luna; grounding review by anthropic/claude-opus-5.5.
Source check: passed by anthropic/claude-opus-5.5 on 2026-11-12T05:03:56.310Z. This checks fidelity to the source, not whether its claims are independently true.
Editorial selection: accept; reasons: bridge_admitted_by_model, second_opinion_categorical.
Observation: 1b85e8db1cd0ccc3c3fd1dd106ab0203ca3946db1e100ecabf8924291df2805b. Source text hash: 1d9a54ca95dc0640d9c9aa3ee253d140e21247dcee74c32b295863f2063aa038.
passed · 2026-11-12T05:03:56.310Z · anthropic/claude-opus-5.5
Publication: 0e61b4c3-c972-4ba1-2140-697b8d80475c
attribution — pass: The Download is a newsletter issue that comments on several items in its own words, so it is one piece. The reading attributes the video-generation claims to the MIT Technology Review newsletter and frames them as the newsletter's characterization rather than as fact.
This is today's edition of The Download, our weekday newsletter that provides a daily dose of what's going on in the world of technology.
summary — pass: The reading accurately conveys the video-generation segment: the named products, the near-indistinguishability claim, and the stated downsides of slop, faked news footage and energy use. It also notes that the underlying methods story is not included. The title and question follow from the evidence. 'AI slop' is quoted from the source and is understandable in context, though a term-library entry for it and for CGI could help.
It’s been a big year for video generation.
date — pass: September 12, 2025 matches the UTC day of originalPublished. 'Last nine months' is kept as the newsletter's own framing. The Deep Dive teasers that mention later events, such as the 22-nation call and OpenAI's math findings, are not used. The modifiedAt timestamp precedes publication, so no later edit affects the cited text. The capture date is disclosed in the limitations.
In the last nine months OpenAI made Sora public, Google DeepMind launched Veo 3, and the video startup Runway launched Gen-4.
support — pass: Each claim matches its excerpt. The reading says spread 'may also make misleading footage more common'; this is hedged and grounded in the source's statement about faked news footage. The energy comparison stays attributed, and the reading notes that it is not quantified.
Video generation also uses up a huge amount of energy, many times more than text or image generation.
media — pass: No media was selected and no verified candidates were supplied, so there is no identity or relevance issue.
Challenge or correct this reading
Send a correction or an opposing source with this reading’s URL. Disagreement and uncertainty can remain unresolved.
OpenAI develops AI models and products and publishes research, release announcements and evaluations.
Explore OpenAIGoogle DeepMind develops AI systems for general tasks and scientific research, including AlphaFold.
Explore Google DeepMindOpenAI develops AI models and products and publishes research, release announcements and evaluations.
Explore OpenAIGoogle DeepMind develops AI systems for general tasks and scientific research, including AlphaFold.
Explore Google DeepMind