The landscape · Media and creative
Media and creative: what AI is really doing
Brief me in 2 minutes
AI in Media and creative.
AI is inside paid media work now, not bolted onto it: a company that sells access to other labs' video models says it hit $1 billion in annualized revenue, Netflix has aired AI-made shots, and Google answers the query before the reader reaches the article. The fight moved from whether training is allowed to who gets paid and how much: labels license the next music model while suing over the last one, and a judge threw out Penske Media's and Chegg's antitrust suits over AI Overviews and said the issue belongs with lawmakers, so what publishers get paid now depends on deals that Google, Perplexity and others choose to offer.
The game: we show you 5 things happening. You call each one real now, early or mostly hype. Then the receipts.
Shift 1 of 5
AI video's money goes to the app that resells the models, not the lab that trains them
Real now
Higgsfield reports a billion-dollar revenue run rate and NVIDIA made its first investment in the company in October. That is paying customers and real capital, not a pilot.
How it works. Runway and Luma train their own video models, and Runway raised $315M in February to pre-train 'world models'. In January Sacra described Higgsfield as a 'model-agnostic aggregator' with 60+ cinematic presets and workflows over models such as Kling, Sora and Veo.
Overstated: Runway's CEO told View Conference in October that AI could 'absolutely' replace human actors, and 'world model' is in every pitch. Higgsfield's billion is an annualized run rate, not a year of booked revenue.
Shift 2 of 5
A judge shuts the antitrust route, leaving publishers with the deals platforms choose to offer
Early
Google itself calls its News AI program a pilot, at more than 400 publications by mid-October; a September 30 report described a pilot, possibly a separate one, paying around 100 publishers for content that helps drive AI answers; and the antitrust route closed at the end of September, subject to appeal.
How it works. AI Overviews answer the query on Google's own page, so the click to the source never happens: Ahrefs found an Overview's presence correlates with a 58% lower click-through rate for the top-ranked page. At the end of September Judge Amit Mehta threw out Penske Media's and Chegg's antitrust suits over those summaries, finding antitrust law does not reach them.
Overstated: 'Google pays publishers' is pilot-stage: the September 30 report said payments ranged from $1,000 over several months for small sites to more than $1 million in a year for one early entrant, and that some larger publishers stayed out to hold out for more. Perplexity's revenue share is real too, and so are the suits against it from Reddit, CNN and others.
Shift 3 of 5
Labels license the next AI music model while suing over the last one
Real now
Warner has a licensing deal with Suno, whose v6 suite has launched, and Universal and Warner have struck licensing deals with Udio; the open question is the liability for the models trained before them, and that is in a courtroom, not a press release.
How it works. Warner switched from suing to licensing: Suno's v6 model suite launched on September 9 with Warner Music Group, BMG and Believe. Universal and Sony are still suing, and their second complaint argues that training a 'new' model on the outputs of an infringing one launders the infringement rather than ending it.
Overstated: Suno says v6 'was trained on content licensed from our partners, interactions including creations and preference signals from our community, and the accumulated learnings from our team'. The labels allege it also learned from earlier models' outputs; Suno's answer does not name their distillation claim, and for that part of the complaint says it 'lacks knowledge or information sufficient to form a belief' and denies the allegations on that basis. No court has ruled, and no receipt here says what an opted-in artist is paid.
Shift 4 of 5
Studios sue over AI training data while writing AI performers into labor contracts instead of banning them
Real now
The SAG-AFTRA contract is ratified, the Midjourney case has reached a stipulated transfer of its video training data, and Netflix has aired the AI shots. None of this is a proposal.
How it works. Disney, NBCUniversal and DreamWorks sued Midjourney in mid-2025 over generated images of Yoda and other owned characters, and in October Midjourney agreed by stipulation to transfer the video datasets it trained on to the studios. In parallel, SAG-AFTRA's ratified contract does not ban synthetic performers: producers may use one only when it adds 'significant additional value', and Netflix has already aired generative AI shots such as a collapsing building.
Overstated: Runway's CEO says AI replacing actors is 'already happening in a way'; Netflix's first aired example was an effects shot, a collapsing building, not a lead performance. Union leaders said when members ratified the contract in June that the 'significant additional value' bar and other provisions would keep the use of AI actors minimal; no receipt here shows how the bar has worked in practice since.
Georgetown Law Technology Institute ↗ChatGPT Is Eating the World ↗
Shift 5 of 5
Consumer AI-video apps rose and died inside one product cycle
Mostly hype
The consumer app OpenAI launched in September 2025 is the one OpenAI itself switched off, app first, API last.
How it works. OpenAI launched Sora 2 on September 30, 2025 as a standalone social app with user 'cameos'. It closed the app on April 26, 2026 and the API on September 24, 2026, and kept Sora only as a research project on world models, while rival models (Google's Veo, Kuaishou's Kling, ByteDance's Seedance) kept shipping inside other companies' apps, the aggregators above included.
Overstated: A capability going mainstream and a specific app winning are two different claims, and Sora's afterlife as a 'world models' research project uses the same phrase as every rival's pitch.
The scary part
What happens to jobs in Media and creative
What AI is already taking:
- generating VFX shots for final footage, like the building collapse in Netflix's 'The Eternaut' TechCrunch ↗
- recording on-camera videos for a business: a personal avatar, now speaking in the person's own voice, makes them without filming each one PR Newswire ↗
- being the answer, so the reader never clicks through to the article that actually reported it Ahrefs ↗
- training new commercial AI music models directly on a major label's catalog, with the label's sign-off Wikipedia ↗
What stays human:
Nobody has shipped an AI system that can direct a two-hour film, run a writers' room, or decide what a publication covers this week; every deal, ruling and lawsuit above is a fight over who owns and gets paid for the training data and the output, not over who makes the creative call. The legal liability for what ships, and the judgment about whether a shot or a track is good enough, still sits with a named human.
Get better · 15 minutes
Owning the parts of the process that carry legal and creative liability: rights clearance, continuity and quality control across an AI-assisted pipeline, and the judgment call on what a synthetic clip or performer is actually cleared to say, since that is exactly what SAG-AFTRA and the studios above are fighting to keep priced and permissioned rather than automated away.
Try this this week:
Rebuild one piece you already made, a training video, a b-roll shot, a jingle, using a named, available tool: HeyGen or Synthesia for a talking-head video, Luma or Runway for a b-roll shot, Suno for a scratch track. Time the whole process including revisions, then compare that hour count and the tool's list price against what the same piece actually cost you last time.
Your move
The model layer (Runway, Luma, Google, Kuaishou) is capital-intensive and already reselling each other: Luma now ships Kling inside its own app. The open space is narrower: rights-cleared data with provenance a buyer can audit, which the Suno laundering fight makes more valuable, and vertical tools that own one workflow end to end, the way Synthesia and HeyGen own corporate talking-head video, rather than another general video app competing with Higgsfield.
Ask whether the company trains its model or resells someone else's: that decides who keeps the margin when the model gets cheaper, and Higgsfield's billion-dollar run rate sits on suppliers it does not control, so probe gross margin and supplier concentration. Red flags: 'world model' claims with no long, controllable output shown, and 'licensed' claims for a model whose predecessor was trained on unlicensed data.
If you publish, pull 90 days of your own analytics this quarter and set Google referral clicks on pages that now show an AI Overview against what AI companies actually pay you (Google's pilot, Perplexity's program or a pay-per-crawl setting at Cloudflare). If the payment column is empty, that gap is your negotiating number, because the antitrust route is shut unless an appeal reopens it.
Prompting is table stakes: Kling now sits inside Luma and Runway alike, and Higgsfield was bundling it in January. What is not commoditized is directing an AI-assisted production: continuity control, rights clearance, and knowing when a synthetic shot or voice needs a signed release. Produce one paid piece end to end with Luma, HeyGen or ElevenLabs and log every hour it actually saved against your last traditional job.
What's actually changing
- 1
AI video's money goes to the app that resells the models, not the lab that trains them
How it works. Runway and Luma train their own video models, and Runway raised $315M in February to pre-train 'world models'. In January Sacra described Higgsfield as a 'model-agnostic aggregator' with 60+ cinematic presets and workflows over models such as Kling, Sora and Veo. Higgsfield says it hit $1 billion in annualized revenue in September, and the labs now resell each other too: Runway added another lab's Wan 3.0 model to its app in August, and in October Luma put Kling 4.0 next to its own tools.
Why it matters. Whoever owns the interface owns the customer and can swap suppliers overnight, while the labs carry the compute bill. When a model lab starts reselling a rival's model too, the model is turning into an input, and the margin is moving to the app.
What's overstatedRunway's CEO told View Conference in October that AI could 'absolutely' replace human actors, and 'world model' is in every pitch. Higgsfield's billion is an annualized run rate, not a year of booked revenue.
HiggsfieldRunwayLuma AIKuaishou (Kling)NVIDIA
4 receipts
- Higgsfield said it reached $1 billion in annualized revenue in September 2026, a month after a $400 million Series B at a $5.4 billion valuation.
That round raised $400 million at a $5.4 billion valuation, and Higgsfield said it hit $1 billion in annualized revenue in September.
Variety, 23 Oct 2026 ↗ - An investor research note describes Higgsfield as an aggregator that bundles other labs' models behind its own presets and workflows.
Higgsfield is a model-agnostic aggregator that bundles Kling alongside Sora and Veo behind 60+ cinematic presets & workflows for quickly generating ads.
Sacra, 30 Jan 2026 ↗ - Luma, which builds its own video models, added Kling 4.0 to its app alongside them.
Kling 4.0 sits next to Luma's own tools and the other video and image options already in your workflow.
Luma AI, 23 Oct 2026 ↗ - Runway made Wan 3.0, a model from another lab, available inside its own app.
Wan 3.0 is now available in Runway in tool mode and workflows.
Runway, 24 Aug 2026 ↗
- Higgsfield said it reached $1 billion in annualized revenue in September 2026, a month after a $400 million Series B at a $5.4 billion valuation.
- 2
A judge shuts the antitrust route, leaving publishers with the deals platforms choose to offer
How it works. AI Overviews answer the query on Google's own page, so the click to the source never happens: Ahrefs found an Overview's presence correlates with a 58% lower click-through rate for the top-ranked page. At the end of September Judge Amit Mehta threw out Penske Media's and Chegg's antitrust suits over those summaries, finding antitrust law does not reach them. What is left is licensing on the buyer's terms: on September 30 PYMNTS, citing The Information, reported that Google had been paying around 100 publishers for how much their content helps drive AI-powered answers, in a pilot involving AI Overviews, AI Mode and the Gemini chatbot, and an executive at one participating publisher said Google seems to determine for itself what that content is worth.
Why it matters. With the antitrust route shut unless an appeal reopens it, publishers get paid only where a platform chooses to pay and at a price the platform appears to set, while the lost clicks hit every site, including the ones with no deal at all.
What's overstated'Google pays publishers' is pilot-stage: the September 30 report said payments ranged from $1,000 over several months for small sites to more than $1 million in a year for one early entrant, and that some larger publishers stayed out to hold out for more. Perplexity's revenue share is real too, and so are the suits against it from Reddit, CNN and others.
GooglePenske MediaCheggPerplexity
7 receipts
- Judge Amit Mehta dismissed Penske Media's antitrust suit over Google's AI summaries, finding antitrust law does not account for them.
U.S. District Judge Amit Mehta found on Wednesday that antitrust laws don't account for AI overviews of content while nodding to "knock on consequences" to publishers whose "content Google takes and repurposes without compensation."
The Hollywood Reporter, 1 Oct 2026 ↗ - A federal judge dismissed Chegg's antitrust suit against Google along with Penske Media's, Reuters reported.
A federal judge in Washington has dismissed antitrust lawsuits that Chegg and Rolling Stone publisher Penske Media Corp. brought against Google, Reuters reported.
Reuters via Northeast Times, 1 Oct 2026 ↗ - Citing a report by The Information, PYMNTS said Google had been paying around 100 publishers for how much their content helps drive AI-powered answers.
Google has reportedly been paying around 100 publishers for how much their content helps drive AI-powered query answers.
PYMNTS, 30 Sept 2026 ↗ - Reported pilot payments range from $1,000 over several months for small sites to more than $1 million in a year for one early participant.
Sources said payments have varied greatly, from $1,000 over the course of several months for smaller sites, while one early entrant in the program has made more than $1 million in a year.
PYMNTS, 30 Sept 2026 ↗ - An AI Overview above a result correlates with a 58% lower click-through rate to the top-ranking page.
the presence of an AI Overview now correlates with a 58% lower average clickthrough rate
Ahrefs, 4 Feb 2026 ↗ - A person close to Google said it began testing the pilot, which involves AI Overviews, AI Mode and the Gemini chatbot, less than a year earlier.
Google started testing the program involving its AI Overviews, AI Mode and Gemini chatbot less than a year ago, said a person close to the company.
PYMNTS, 30 Sept 2026 ↗ - Google said more than 400 publications worldwide were participating in its News AI pilot program.
Since launching our News AI pilot program last year with a range of news partners, like The Washington Examiner and The Washington Post, more than 400 publications worldwide are now participating in this program.
Google, 15 Oct 2026 ↗
- Judge Amit Mehta dismissed Penske Media's antitrust suit over Google's AI summaries, finding antitrust law does not account for them.
- 3
Labels license the next AI music model while suing over the last one
How it works. Warner switched from suing to licensing: Suno's v6 model suite launched on September 9 with Warner Music Group, BMG and Believe. Universal and Sony are still suing, and their second complaint argues that training a 'new' model on the outputs of an infringing one launders the infringement rather than ending it. In its October answer Suno admitted it has not deleted all of the recordings it used to train its models.
Why it matters. A licence on the next model does not settle the last one: if a court accepts the laundering theory, a 'licensed' model built on an unlicensed predecessor carries the old liability with it, and so does every product built on top.
What's overstatedSuno says v6 'was trained on content licensed from our partners, interactions including creations and preference signals from our community, and the accumulated learnings from our team'. The labels allege it also learned from earlier models' outputs; Suno's answer does not name their distillation claim, and for that part of the complaint says it 'lacks knowledge or information sufficient to form a belief' and denies the allegations on that basis. No court has ruled, and no receipt here says what an opted-in artist is paid.
SunoUdioElevenLabsWarner Music GroupUniversal Music GroupSony Music
3 receipts
- Suno launched its v6 model suite on September 9, 2026 with Warner Music Group, BMG and Believe.
The labels' central new allegation concerns v6, the model suite Suno launched on September 9 with Warner Music Group, BMG, and Believe.
Music Business Worldwide, 22 Oct 2026 ↗ - Universal and Sony allege that training v6 on the outputs of earlier models launders the infringement.
Training a 'new' model on the outputs of an infringing model does not eliminate the infringement; it launders it.
Music Business Worldwide, 22 Oct 2026 ↗ - In its answer to the second suit, Suno admitted it has not deleted all of the recordings it trained on.
Suno has admitted that it has not deleted all of the sound recordings it used to train its AI models.
Music Business Worldwide, 22 Oct 2026 ↗
- Suno launched its v6 model suite on September 9, 2026 with Warner Music Group, BMG and Believe.
- 4
Studios sue over AI training data while writing AI performers into labor contracts instead of banning them
How it works. Disney, NBCUniversal and DreamWorks sued Midjourney in mid-2025 over generated images of Yoda and other owned characters, and in October Midjourney agreed by stipulation to transfer the video datasets it trained on to the studios. In parallel, SAG-AFTRA's ratified contract does not ban synthetic performers: producers may use one only when it adds 'significant additional value', and Netflix has already aired generative AI shots such as a collapsing building.
Why it matters. Studios are running two tracks at once: litigate the input, what the models were trained on, while normalizing the output under negotiated rules. Midjourney's video training data is now moving to the plaintiffs, the evidence every later case will want.
What's overstatedRunway's CEO says AI replacing actors is 'already happening in a way'; Netflix's first aired example was an effects shot, a collapsing building, not a lead performance. Union leaders said when members ratified the contract in June that the 'significant additional value' bar and other provisions would keep the use of AI actors minimal; no receipt here shows how the bar has worked in practice since.
DisneyNBCUniversalDreamWorksMidjourneySAG-AFTRANetflixRunway
4 receipts
- Disney, NBCUniversal and DreamWorks sued Midjourney in mid-2025, showing it reproducing characters like Yoda from simple prompts.
the AI is willing to produce fairly realistic depictions of Star Wars character Yoda in response to the prompt
Georgetown Law Technology Institute, 24 June 2025 ↗ - Midjourney agreed by stipulation to transfer the video datasets it trained on to the studio plaintiffs.
By stipulation of the parties in Disney v. Midjourney, Midjourney will be transferring its video datasets used to train its models to the Plaintiffs.
ChatGPT Is Eating the World, 7 Oct 2026 ↗ - SAG-AFTRA members ratified a four-year contract under which AI performers must bring significant additional value over a live actor before producers can use them.
The contract says AI performers must bring “significant additional value” over a live actor or a digital capture of them if producers are to use them.
Associated Press (WTOP), 5 June 2026 ↗ - Asked whether AI could replace human actors, Runway's co-CEO said yes, and that it is already happening.
"Yeah, absolutely. I think it's already happening in a way," he replied, going on to describe AI as "a new medium of expression" and "a new medium of storytelling."
Variety, 20 Oct 2026 ↗
- Disney, NBCUniversal and DreamWorks sued Midjourney in mid-2025, showing it reproducing characters like Yoda from simple prompts.
- 5
Consumer AI-video apps rose and died inside one product cycle
How it works. OpenAI launched Sora 2 on September 30, 2025 as a standalone social app with user 'cameos'. It closed the app on April 26, 2026 and the API on September 24, 2026, and kept Sora only as a research project on world models, while rival models (Google's Veo, Kuaishou's Kling, ByteDance's Seedance) kept shipping inside other companies' apps, the aggregators above included.
Why it matters. OpenAI shut its standalone app first and its API last, while rival models kept shipping inside other companies' apps. OpenAI gave no specific reason in its shutdown notice.
What's overstatedA capability going mainstream and a specific app winning are two different claims, and Sora's afterlife as a 'world models' research project uses the same phrase as every rival's pitch.
OpenAIGoogle DeepMindKuaishouByteDance
5 receipts
- OpenAI unveiled Sora 2 with an iOS app on September 30, 2025.
Sora 2 was unveiled on September 30, 2025, with an iOS app at the same time
Wikipedia, 30 Sept 2025 ↗ - OpenAI set a two-stage shutdown: the app on April 26, 2026 and the API on September 24, 2026.
The web and app version goes dark on April 26, 2026, with the Sora API following on September 24, 2026.
The Decoder, 28 Mar 2026 ↗ - Both stages happened on schedule.
The Sora app was shut down on April 26, 2026, while the API was shut down on September 24, 2026.
Wikipedia, 24 Sept 2026 ↗ - OpenAI gave no specific reason for discontinuing Sora in its shutdown notice.
OpenAI did not provide a specific reason for discontinuing Sora in its shutdown notice.
Wikipedia, 24 Mar 2026 ↗ - OpenAI's own notice says the Sora web and app experiences were discontinued on April 26, 2026, and states the API shutdown date as September 24, 2026.
The Sora web and app experiences were discontinued on April 26, 2026. The Sora API will be discontinued on September 24, 2026.
OpenAI Help Center, 20 Oct 2026 ↗
- OpenAI unveiled Sora 2 with an iOS app on September 30, 2025.
Who's doing it
11 companies to know, 7 with a round we can source. Newest first where dated.
Search these companies in the startup database ↗ · Explore all research tools
Higgsfieldvideo and image generation, model aggregator
AI video and image platform for creators, brands and agencies that Sacra described in January as a 'model-agnostic aggregator', bundling rivals' models such as Kling and Veo behind presets and workflows.
Latest · 23 Oct 2026Extended its Series B with new money from NVIDIA and TCV, after saying it hit $1 billion in annualized revenue in September. Variety ↗
$400M Series B, 17 Aug 2026 source ↗
Luma AIvideo generation, creative app
Builds its own video models and a board-based creative app, which now also hosts rival labs' models.
Latest · 23 Oct 2026Added Kling 4.0, a rival lab's video model, to its own app next to Luma's own tools. Luma AI ↗
$900M Series C, 20 Nov 2025 source ↗
SunoAI music generation
Generates full songs, including vocals, from text prompts; licensed by Warner for its v6 models while Universal and Sony still sue.
Latest · 22 Oct 2026Admitted in its answer to Universal and Sony's second lawsuit that it has not deleted all of the recordings it used to train its models. Music Business Worldwide ↗
$400M Series D, 3 June 2026 source ↗
Black Forest LabsAI image and video models
Makes the FLUX family of image and video models, licensed into creative tools from Adobe, Canva, Meta and Microsoft and also released as open weights.
Latest · 15 Oct 2026Released open weights for FLUX 3 Image Klein, a 5.6 billion parameter image model small enough to run on consumer GPUs and laptops. GlobeNewswire ↗
$300M Series B, 1 Dec 2025 source ↗
HeyGenAI avatar video
AI avatar video platform that generates talking-head presenter video for marketing and training without booking a person or a shoot, now with its own voice model.
Latest · 9 Oct 2026Launched HeyGen Voice, its own in-house voice model, which it says puts identity and voice in one place rather than a patchwork of vendors. PR Newswire ↗
MidjourneyAI image generation
Text-to-image and video generator, sued by Disney, NBCUniversal and DreamWorks over copyright, and handing its video training datasets to them by stipulation.
Latest · 7 Oct 2026Agreed by stipulation to hand the video datasets it trained on to Disney and the other studio plaintiffs. ChatGPT Is Eating the World ↗
ElevenLabsvoice generation and cloning
Voice generation, dubbing and voice-agent platform, also behind the ElevenMusic song generator; in September it signed Universal Music Group, its first deal with a major record company.
Latest · 1 Oct 2026Completed a $300M employee tender offer at a $22B valuation, double its February Series D price; the money went to the selling staff, not the company. Music Business Worldwide ↗
$500M Series D, 4 Feb 2026 source ↗
Runwayvideo generation, world models
Trains its own text- and image-to-video models, raised $315M in February to pre-train 'world models' for media, gaming and robotics, and now sells an ad agent to brands.
Latest · 30 Sept 2026Launched Runway Ads, an agent that connects to a brand's Meta, Google and TikTok accounts, makes and localizes ad variations, and pulls the performance data back. Adweek ↗
$315M Series E, 10 Feb 2026 source ↗
SynthesiaAI avatar video, corporate training
AI avatar video platform enterprises use for training and product videos without a camera crew, now adding live two-way avatar conversations.
Latest · 28 Sept 2026Launched Sessions, two-way video conversations with AI avatars for roleplay practice and survey-style interviews, moving beyond one-way training video. Synthesia ↗
$200M Series E, 26 Jan 2026 source ↗
PerplexityAI search, publisher licensing
AI answer engine and Comet browser that answers queries directly from indexed content, shares revenue with publishers in its program, and has been sued by others.
Latest · 3 Aug 2026A federal judge largely rejected its bid to dismiss Reddit's suit, finding Reddit plausibly pleaded that Perplexity conspired with a data scraper to get around Reddit's access controls. Press Gazette ↗
UdioAI music generation
AI song generator that settled with Universal and Warner and struck licensing deals with them; when Sony sued it a second time in July, Music Business Worldwide reported that Sony was the only major yet to reach a licensing agreement with it.
Latest · 20 July 2026Sony sued it a second time, over 30,117 recordings it says Udio copied to train its models, while Universal and Warner have already settled and licensed it. Music Business Worldwide ↗
Your move
Our read, built on the receipts above.
The model layer (Runway, Luma, Google, Kuaishou) is capital-intensive and already reselling each other: Luma now ships Kling inside its own app. The open space is narrower: rights-cleared data with provenance a buyer can audit, which the Suno laundering fight makes more valuable, and vertical tools that own one workflow end to end, the way Synthesia and HeyGen own corporate talking-head video, rather than another general video app competing with Higgsfield.
Ask whether the company trains its model or resells someone else's: that decides who keeps the margin when the model gets cheaper, and Higgsfield's billion-dollar run rate sits on suppliers it does not control, so probe gross margin and supplier concentration. Red flags: 'world model' claims with no long, controllable output shown, and 'licensed' claims for a model whose predecessor was trained on unlicensed data.
If you publish, pull 90 days of your own analytics this quarter and set Google referral clicks on pages that now show an AI Overview against what AI companies actually pay you (Google's pilot, Perplexity's program or a pay-per-crawl setting at Cloudflare). If the payment column is empty, that gap is your negotiating number, because the antitrust route is shut unless an appeal reopens it.
Prompting is table stakes: Kling now sits inside Luma and Runway alike, and Higgsfield was bundling it in January. What is not commoditized is directing an AI-assisted production: continuity control, rights clearance, and knowing when a synthetic shot or voice needs a signed release. Produce one paid piece end to end with Luma, HeyGen or ElevenLabs and log every hour it actually saved against your last traditional job.
What's next
- 3 Nov 2026 Suno must answer the class action brought by Jason Isbell and other musicians over artist-name prompts. musicbusinessworldwide.com ↗
- 20 Nov 2026 Briefing is scheduled to close on the fair-use summary judgment motions in The New York Times v. OpenAI and Microsoft; as of early October no trial date had been set. openclassactions.com ↗
Other industries
- FinanceBanks move agentic AI from chatbots to running back-office and dealmaking work
- Software engineeringSpaceX owns Cursor and starts rewiring it: one Grok bundle, and OpenAI's models on the way out
- HealthcareAmbient scribes go from pilot project to clinical infrastructure
- LegalLegal AI vendors go from research assistant to agent that produces the deliverable
- Real estateLeasing AI graduates from answering the phone to acting inside the property system
- Sales and marketingCustomers start sending their own AI agents, and vendors race to build the front door for them
- Jobs, HR and recruitingAI runs the first-round interview, not just the resume filter
- InsuranceClaims move from adjuster-first to AI-first, humans supervise the exceptions
- CybersecurityAttackers hand whole intrusion campaigns to AI agents, at about $50 a breach
- Consulting and accountingAI-written client reports become a recurring, public Big Four embarrassment
- Retail and e-commerceAgent checkout retreats to a cart, and the banks guard the last tap
- EducationNYC and LA ban AI for students, not for teachers
- Manufacturing and roboticsHumanoid robots move from demo livestreams to paid warehouse shifts
- Transport and robotaxisRobotaxis go from demo to metered fare, and Waymo starts borrowing to scale
- Energy and data-centre powerGrid rules get rewritten mid-buildout, state by state
- GovernmentGSA turns AI adoption into a government-wide subscription, then swaps 'free' for metered contracts
- Agriculture and foodDriverless tractors learn to run the tools behind them, through open platforms and retrofit kits