The AI landscape doesn't move in one direction — it lurches. Some techniques leap from experiment to table stakes in a single quarter; others stall against regulatory walls, technical ceilings, or organisational inertia that no amount of hype can dislodge. Knowing which is which is the hard part. The State of Play cuts through the noise with a rigorously maintained index of AI techniques across every major business domain — classified by maturity, evidenced by real-world adoption, and updated daily so you always know where you stand relative to the field. Stop guessing. Start knowing.
A daily newsletter distilling the past two weeks of movement in a domain or two — delivered to your inbox while the index updates in the background.
Each dot marks the weighted maturity of practices within a domain — hover for a brief summary, click for more detail
AI for generating and editing images, video, audio, 3D assets, and cross-media content. Mostly leading-edge with rapid advancement — image generation, music composition, and voice synthesis are approaching good practice. Video generation and 3D asset creation are progressing fast but quality and controllability gaps persist. The most active domain by momentum: over half the practices are advancing.
The headline: The AI tools for making images, video, voice, and music now work brilliantly — but customers, courts, and fraud have become the real barriers, not the technology. The winners are companies using AI for narrow, low-stakes tasks; everyone chasing open-ended "AI creativity" is hitting a wall.
Most companies in creative work have access to AI that genuinely matches human output on simple tasks, and a small group is scaling it profitably — but only in tight, repeatable lanes like sports highlight reels, product mockups, and training videos. The wider race to use AI for headline creative work has stalled across the board, and the reason has nothing to do with the technology. Audiences are turning against it: consumer acceptance of AI-made content has fallen from roughly 60 percent in 2023 to the mid-20s today, and 83 percent of consumers say they can spot it and trust the brand less when they do. Meanwhile the courts are catching up — major studios are suing the image-generation tools, and new disclosure laws carry real fines. If you are deploying AI in marketing or media, the question has flipped from "is it good enough?" to "will customers and regulators accept it?"
Voice cloning crossed from "promising" to "stuck." The technology is now flawless — a clone needs only a few seconds of someone's voice and people cannot tell the difference over a phone. That very perfection is the problem: a US federal bill (the NO FAKES Act) advanced with penalties up to $750,000, and voice-scam fraud is up more than tenfold in a year. Treat any voice-AI rollout as a legal and fraud-control project, not a creative one.
The leading deepfake expert says detection is broken. Hany Farid, the field's most cited authority, publicly admitted he can no longer tell real from fake. The industry's answer is now "provenance" — invisible tags proving where content came from — but researchers just showed those tags can be stripped off. Do not rely on detection software to protect your brand or verify content; assume anything unsigned could be fake.
The video-game industry is openly rejecting AI. Take-Two confirmed its blockbuster Grand Theft Auto 6 uses no generative AI, and games that disclose AI use are getting 53 percent fewer reviews — a sales warning sign. Where your audience cares about craft, visible AI use can actively cost you customers.
The bright spot: narrow automation is paying off. Adobe shipped AI that auto-assembles video edits, and live AI highlight clips ran across the World Cup within seconds of each goal. Enterprise training videos now cost roughly $8 versus $3,000+ the old way. The reliable returns are in bounded, repetitive tasks — start there.
August 2: EU labeling rules go live. The EU AI Act starts requiring clear labels on AI-generated and manipulated content, with fines up to 3 percent of global revenue. If you sell or advertise in Europe, audit now where AI touches your customer-facing content and make sure it can be disclosed.
July 31: a German court rules on AI music training. The GEMA v. Suno decision will be an early test of whether training AI on copyrighted work is infringement — a signal for every company relying on AI-generated images, music, or copy. Watch it; it shapes your legal exposure.
Disclosure is becoming a competitive edge, not just a cost. Brands that proactively label AI use are seeing higher trust and conversion in early tests, while consumers increasingly punish those who hide it. Build disclosure into your content process now rather than retrofitting it under regulatory pressure.
The audience is the constraint you can't engineer around. Better models won't fix the fact that customers increasingly distrust content once they realize it's AI — H&M's AI-model product pages converted 22 percent worse than human ones. The limit is acceptance, not quality.
The same tools that create also defraud. Voice clones and deepfakes are both your shiny new capability and an industrialized crime wave — one bank logged over 8,000 deepfake fraud attempts in a single quarter. Adopting these tools means owning the fraud-and-consent risk that comes with them.
A great model is not a viable business. OpenAI shut down its Sora consumer video app despite stunning output because it burned about $1 million a day against almost no revenue. Impressive demos don't pay; bounded, repeatable use cases do.
Go deeper: the full Creative & Generative Media briefing — the longer analytical write-up, plus every practice we track in this domain with its maturity rating, the tools to consider, and the evidence behind our assessment.