Your brand deserves a shot, make it a bold one with Takila Design.

    What makes us different? A rare mix of deep expertise and a burning passion for design.

    How Creative Directors Are Beating AI Every Day

    AI makes the options. Creative directors make the difference.

    In a 2025 experiment published in Nature Human Behaviour, researchers asked people to invent a toy using a brick and a fan. The people working with ChatGPT kept landing in the same place: 94% of their ideas shared overlapping concepts, and nine of them independently named their toy the “Build-a-Breeze Castle.” The people working without AI came up with ideas that were entirely unique (Wharton Mack Institute).

    That toy name sums up the creative director’s job in 2026.

    We use AI every day. It’s in our moodboards, our previs and our 3D-to-AI production pipeline, and it makes us faster. So when we say creative directors are beating AI, we don’t mean refusing it. We mean something more specific: making work that ends up better than anything the machine would produce on its own.

    First, the scoreboard, because it isn’t one-sided. AI wins on volume, speed and the average idea. In a Wharton-led study revised in July 2026, AI-generated product ideas were seven times more likely to rank in the top 10% for purchase intent (RePEc). In a 2024 Science Advances study, AI helped the least creative writers most, lifting their novelty by up to 10.7% and their usefulness by 11.5%, while the most creative writers gained little (ScienceDaily).

    AI raises the floor. The question is who raises the ceiling. Here are five things creative directors do every day that AI can’t do for them.

    1. They fight sameness

    The same research that credits AI on average quality also shows it narrows the range. In the brainstorming studies, AI-assisted ideas were less diverse in 37 of 45 comparisons. In the Science Advances study, stories written with AI help were more similar to one another. The authors call this a “social dilemma”: each writer gains while the pool of fresh ideas shrinks. A 2025 meta-analysis of 28 studies and 8,214 participants found no overall creativity gap between humans and AI, a modest gain when humans work with AI, and a large drop in the diversity of ideas (arXiv).

    In advertising, that is the core problem. When every brand and studio uses the same models, default output converges, and convergence is the opposite of a brand.

    The fixes that work in the research are, tellingly, creative-direction techniques applied to a machine: adding constraints and personas, pooling ideas from different sources, and deliberately exploring away from the obvious (RePEc). Some creatives now run a simple originality test. If an AI tool comes up with the same idea when prompted, they throw it out and start again (Adobe). Every morning, the job is to make sure the work isn’t another Build-a-Breeze Castle.

    2. They write the brief and find the insight

    AI is very good at answers. Deciding which question is worth asking is still a human job.

    Kantar’s November 2025 analysis shows what’s at stake. AI-generated ads that don’t look obviously artificial can perform well: more than 40% reached the top tier for branded cut-through. On average, though, they scored lower on branding, mostly where the AI wasn’t tailored to the brand’s tone and assets. An unofficial AI ad for Liquid Death drew strong emotion but ranked in the bottom 25% for brand memorability, because the brand was missing from its most engaging moments (Kantar). NIQ’s neuroscience testing found that even AI ads viewers rated as high quality produced weaker memory activation than traditional ads (NIQ).

    A model can produce a striking moment. Knowing where the brand belongs inside that moment, and why this audience should care, is the brief. That comes from client conversations, category knowledge and a point of view. It’s why India’s senior creatives describe clients wanting their creative leaders hands-on, shaping ideas and challenging briefs (Storyboard18).

    3. They choose, and they commit

    Cheap generation brings a new problem: too many options. In MikeTeevee’s Brand Perspectives 2026, HelloFresh ECD James Hurst said, “Some teams are suffering from real decision paralysis.” Wayfair’s head of creative asked how to stop everything looking the same (Creative Boom).

    That makes choosing the scarce skill. David Droga said it plainly: “Taste is a real thing” (Wharton AI & Analytics). Rediffusion’s Carol Goyal put it more sharply: “AI has democratised generation. It has not democratised judgement” (Storyboard18). Adobe’s June 2026 study of 1,433 US and UK creatives found human work shifting towards “selection, direction, client negotiation, and accountability for how work was made” (Adobe).

    Day to day, it looks like this. A hundred AI frames arrive in an afternoon. The creative director picks one, defends it in the client room, then crafts it until it’s finished. Conviction is what turns a variant into a campaign.

    4. They read India from the inside

    Generative models mostly learn from someone else’s internet. Researchers from Penn State and the University of Washington worked with Indian communities and found that text-to-image models portray Indian cultures “through an outsider’s lens.” Celebrations come out more colourful than they really are, some subcultures are over-represented, and exotic stereotypes get reinforced. The researchers note that the errors look accurate at first glance, which makes them easy to miss (Penn State IST). Researchers at IIT Jodhpur and Meta built a separate benchmark across 30 Indic languages, because image generation quality and cultural relevance outside English need their own testing (arXiv).

    This is a commercial problem as much as a cultural one. LS Digital reports that vernacular content strategies deliver 30–40% higher engagement, yet only 48% of Indian users trust AI-generated content, and about 73% of marketers say AI should augment people, with human oversight, rather than automate them (Storyboard18).

    A Diwali film, a regional-language cutdown, a wedding-season launch: each depends on details a model gets nearly right. And “nearly right” is how a brand ends up looking like a tourist in its own market. Catching that takes someone who knows the culture from the inside.

    5. They put their name to the work

    AI can’t be held accountable. Someone has to be.

    The industry learned this publicly in June 2025. Cannes Lions withdrew DM9’s Creative Data Grand Prix after AI-generated and manipulated content in a case film gave the jury inaccurate information, then made AI disclosure mandatory for entries (Cannes Lions). For 2026 the festival added AI Craft subcategories. Entrants have to show where human creativity guided the outcome and whether AI made the work better, “not just faster” (Campaigns of the World).

    India’s rules have moved too. MeitY’s IT Amendment Rules, in force since 20 February 2026, require platforms to label synthetically generated content prominently and embed provenance metadata (Business Today). ASCI’s final guidelines on AI in advertising, issued on 29 September 2026, prohibit three things: deepfakes or likenesses used without consent, fabricated testimonials, and misleading synthetic depictions of product results. They also require a label wherever synthetic content materially influences a buying decision, for example a demo of a product that doesn’t exist yet (Fox Mandal).

    Disclosure has a cost. Researchers at NYU Stern found that fully AI-generated visual ads could beat ads by human experts on click-through by up to 19%, but telling viewers an ad was AI-made cut click-through by 31.5% (NYU Stern). So someone has to decide where synthetic imagery belongs, where real product accuracy has to carry the frame, and how to stay honest with the audience without weakening the work. That takes a person who understands the brand, the rules and the viewer, and who will answer for the result.

    AI makes options. Creative directors make the work.

    None of this argues against AI. Droga has described its real gift as getting you to bad ideas faster, so you can move past them sooner. The research points the same way. The strongest results come from people directing machines, not from either working alone.

    That’s how we work at Takila Design. AI widens the field: more references, faster previs, more variations. Creative direction narrows it down to the idea, the frame and the finish that could only belong to one brand. (For the production side of that pipeline, read our companion piece, Why 3D Production Is Still Relevant in the Age of AI.)

    We’ve crafted 3D product films, CGI and brand work for names such as Lava, Somany, Casio, Kajaria and Citizen. If you want work that uses everything AI can do and still looks like nobody else’s, we’d love to talk. See our work at takiladesign.com or write to us at cheers@takiladesign.com.


    Sources

    1. Wharton Mack Institute, summary of Meincke, Nave & Terwiesch, “ChatGPT decreases idea diversity in brainstorming,” Nature Human Behaviour (14 May 2025). https://mackinstitute.wharton.upenn.edu/2025/new-in-nature-chatgpt-decreases-idea-diversity-in-brainstorming/
    2. Terwiesch, Meincke, Girotra, Mollick, Nave & Ulrich, “AI and Its Impact on Creativity and Diversity: An Empirical Study of LLM-Generated Product Ideas,” arXiv 2607.27553 (revised Jul 2026), via RePEc/IDEAS. https://ideas.repec.org/p/arx/papers/2607.27553.html
    3. ScienceDaily, summary of Doshi & Hauser, “Generative AI enhances individual creativity but reduces the collective diversity of novel content,” Science Advances 10(28) (Jul 2024). https://sciencedaily.com/releases/2024/07/240712222127.htm
    4. Holzner, Maier & Feuerriegel, “Generative AI and Creativity: A Systematic Literature Review and Meta-Analysis,” arXiv 2505.17241 (22 May 2025). https://arxiv.org/abs/2505.17241
    5. Adobe Research, “How AI is redistributing creative work” (Jun 2026). https://www.adobe.com/ai/research/202606/ai-is-redistributing-creative-work.html
    6. Kantar, “Rethinking AI-Generated advertising” (4 Nov 2025). https://www.kantar.com/north-america/Inspiration/Advertising-Media/Rethinking-AI-Generated-advertising
    7. NIQ, “NIQ research uncovers hidden consumer attitudes toward AI-generated ads” (12 Dec 2024). https://nielseniq.com/global/en/news-center/2024/niq-research-uncovers-hidden-consumer-attitudes-toward-ai-generated-ads/
    8. Storyboard18, “From storytellers to system builders: what happened to CCOs in Indian advertising” (undated). https://www.storyboard18.com/brand-makers/from-storytellers-to-system-builders-what-happened-to-ccos-in-indian-advertising-82848.htm
    9. Creative Boom, MikeTeevee Brand Perspectives 2026: “Decision paralysis, lost conviction and no juniors” (21 Sep 2026). https://www.creativeboom.com/state-of-play/decision-paralysis-lost-conviction-and-no-juniors-what-brand-creative-leaders-say-ai-is-exposing/
    10. Wharton AI & Analytics, “AI, taste and the future of creativity: David Droga on what’s worth saving” (22 Jul 2025). https://ai.wharton.upenn.edu/updates/ai-taste-and-the-future-of-creativity-david-droga-on-whats-worth-saving/
    11. Storyboard18, “The AI problem in agencies: why originality is becoming a premium skill again” (c. 2026). https://www.storyboard18.com/advertising/the-ai-problem-in-agencies-why-originality-is-becoming-a-premium-skill-again-87382.htm
    12. Penn State College of IST, on Venkit et al.’s study of text-to-image models and Indian cultures (AAAI/ACM AIES, c. Oct 2024). https://www.ist.psu.edu/about/news/venkit-aies-social-harms
    13. Mittal et al., IndicTTI benchmark, arXiv 2408.00283 (Aug 2024; ECCV 2024). https://arxiv.org/abs/2408.00283
    14. Storyboard18, LS Digital report on AI-native advertising in India (11 Dec 2025). https://www.storyboard18.com/advertising/ai-native-advertising-surges-in-india-as-report-flags-huge-knowledge-gap-among-consumers-85684.htm
    15. Cannes Lions, official statement on DM9 entries (27 Jun 2025). https://www.canneslions.com/news/cannes-lions-statement-dm9-entries-into-cannes-lions-2025
    16. Campaigns of the World, Cannes Lions 2026 AI Craft subcategory (1 Jun 2026). https://campaignsoftheworld.com/news/ai-craft-subcategory-cannes-lions-2026/
    17. Business Today, “Govt tightens rules on AI-generated content, mandates labels and faster takedowns” (10 Feb 2026). https://www.businesstoday.in/technology/story/govt-tightens-rules-on-ai-generated-content-mandates-labels-and-faster-takedowns-515505-2026-02-10
    18. Fox Mandal, “ASCI issues guidelines for labelling of AI content in advertising” (Sep/Oct 2026). https://foxmandal.in/news/asci-issues-guidelines-for-labelling-of-ai-content-in-advertising/
    19. NYU Stern, Ghose, Lee, Todri & Adamopoulos, “The AI Advertising Paradox” (6 Nov 2025). https://www.stern.nyu.edu/experience-stern/faculty-research/ai-advertising-paradox