Why Kolsquare is publishing its AI Manifesto, what’s the impact, and what we still have to figure out.

AI is changing influencer marketing faster than anything since social platforms arrived. Most of it is good news. Some of it worries me, and I suspect I’m not the only one. At Kolsquare we’re part of it, because we analyses the public content of millions of creators and sells AI features. But today I am publishing our Kolsquare AI Manifesto: what we mean by « better AI », six things we won’t do, and five commitments we’ll report on every year. Some of these commitments will cost us: business might turn down, features that are hard to build, numbers that may not look good. It’s a first version, and we will get some things wrong along the way, but I would love help getting them right.

This summer, I stood in front of my whole Kolsquare team and put one question on the screen: what kind of creator economy do we want AI to build?

I didn’t have the answer. I still don’t have all of it. But I’ve spent the past months working on it with our leadership team, our product and tech teams, our mission committee and a lot of people outside Kolsquare. Today, as Paris Creator Week 2026 opens, I am publishing what came out of it.

This is a first version. Some of what we commit to has never been built in our industry, and I don’t know yet exactly how we’ll get there. But we will try things, get some of them wrong, and adjust. Because we feel like we need to.

And I want to explain why, because the « why » matters more to me than the document itself.

The best and the most worrying thing to happen to our industry in years

I use AI every day. So does almost everyone I work with. In our own research, 84% of influencer marketers say they already use AI or automation tools, and so do 72% of creators [Source: Kolsquare, The Future of Influencer Marketing 2026-2027 and Voices of the Creator Economy].

The benefits are real. Brands find new creators in minutes instead of days. Agencies spend their time on strategy instead of spreadsheets. Creators edit, subtitle and translate their content in languages they don’t speak. And something new is happening: when people ask ChatGPT, Gemini or Perplexity what to buy, the answers often draw on what creators have said. AI is becoming a second audience for creator content, alongside people.

But I see the other side every week too.

Feeds are filling up with synthetic content, and audiences struggle to tell what’s real. Virtual influencers compete with people for brand budgets. Bots answer comments and DMs in creators’ place. Agents now contact and negotiate with creators on behalf of brands, without always saying they’re machines. Every AI query uses energy and water, and relies on people (often underpaid) who label data and moderate content out of sight.

And the money isn’t there yet for most people. In a Bain survey published in September, 95% of marketing and finance executives said their company uses AI tools, and only 6% said it delivers substantial results [Bain via Business Insider, Sept 2026]. So we’re running fast, without always knowing where.

To me, the real risk is technology built without purpose and without responsibility.

We’re part of this

I have a conflict of interest, and I want to be clear about it. At Kolsquare, we sell software. We use machine learning, computer vision, language and image analysis, and generative AI. We’re building agents. I’m writing about the risks of something we make money from.

There’s more. Influencer marketing lives on attention, and attention is exactly what these systems are built to capture. Kolsquare analyses the public content of millions of creators, and our tools shape which of them brands get to see. That’s potentially a lot of influence on people’s livelihoods.

I could pretend we stand outside the problem. I’d rather say where we stand.

Why a manifesto, and why now

This isn’t the first time we’ve had to make this kind of choice at Kolsquare.

We’re far from perfect, but we’ve been pioneering responsible influence for years. When we started, many people thought it was a nice idea that would slow us down. I felt a bit crazy for years. We became an Entreprise à Mission in 2020 and a B Corp in 2024, first of our industry. We stated our vision and values in our Kolsquare Culture Book. We turned down clients when they went against our mission (ultra-fast fashion brands are one example). And we give 1% of our revenue every year to NGOs working on the limits of social media and influence (harassment, fake news, mental health, environmental impact, etc.).

What I learned from all of it: values only become real at the point where they cost you something. Our own first drafts were exactly that: 10 exciting principles about « trustworthy AI » that cost nothing. It took us months, and a few uncomfortable discussions, to turn them into choices.

So we tried to write the opposite: choices we’d have to defend, including in front of clients who’ll ask for things we won’t build.

Here it is, in full.

---

THE KOLSQUARE AI MANIFESTO

Version 1, November 2026.

WHAT’S HAPPENING TODAY IN THE CREATOR ECONOMY WITH AI

AI is changing influencer marketing faster than anything since the arrival of social platforms. Brands find creators in minutes. Agencies build reports in a few clicks. Creators edit, subtitle and translate with tools that didn’t exist two years ago. And AI agents now brief, contact and negotiate with creators on behalf of brands.

The whole marketing chain is shifting with AI. People now compare products by asking ChatGPT, Gemini or Perplexity, and the answers they get often draw on what creators have said. After SEO, creators now shape GEO. AI is becoming a second audience for their content, alongside people.

Some of this is great news. Some of it worries us.

Feeds are filling up with synthetic and inauthentic content, and audiences struggle to tell what’s real. Virtual influencers compete with people for brand budgets. Machines are taking over everyday interactions too: bots reply to comments and DMs in creators’ place, chatbots have replaced the people you used to reach to report a problem on a platform, and AI-run accounts pass as real followers. Creators get sourced, scored and negotiated with by machines, often without knowing it. Every AI query uses energy and water, and relies on people, often underpaid, who label data and moderate content out of sight.

We’re part of this. Influencer marketing lives on attention, and attention is what these systems are built to capture. Kolsquare analyses the public content of millions of creators, and our tools shape which of them brands get to see. That’s a responsibility. We’ve used machine learning, computer vision, language and image analysis for years, and we use generative AI every day, in our product and in our own work. We also see our users and all market players, from creators to agencies and brands, using more and more AI to better perform in their campaigns and streamline their processes.

So we won’t pretend to stand outside the problem. But we’d rather say where we stand, and what we’re going to do about it.

WE BELIEVE IN BETTER AI (NOT BETTER AI)

Our industry is in an AI race. Every week brings a new agent, a new connector, a new copilot, a new « first ». We will keep building fast, because that’s our job. But we will judge each step with one question: does it make influence better?

Our answer starts with what we mean by better AI.

Better AI serves the people who make influence work. Creators first, because AI should never take a real human creator’s place and business. Then talent managers, agencies and brand teams, whose judgment and relationships make the magic of campaigns work.

Better AI is visible. Brands, creators and audiences should know when AI is involved, understand why it produced a result, and be able to switch generative AI off.

Better AI is as light as possible. AI always has a footprint. We measure ours, and work every year to reduce it.

Better AI is fairer. AI can open influence to more creators and smaller brands, or concentrate value on the biggest accounts and the biggest platforms. We want ours to widen the field. It’s what Kolsquare has worked on since day one, growing the market so more brands can reach their potential with creators, and so that more creators can make a living from what they love.

OUR ETHICAL STANDARDS FOR AI IN INFLUENCER MARKETING

At Kolsquare we’ve turned down business before when it went against our mission (ultra-fast fashion brands are one example). AI in influencer marketing deserves the same discipline.

We won’t build tools that clone creators, even with their consent. We won’t build tools that copy a creator’s content or style without their consent. Nor will we build tools that promote AI influencers or steer brands towards them. Virtual influencers will probably keep a small place in the market. We won’t help them take work from real people.

We won’t let our AI decide on its own who gets a collaboration. It can recommend, rank and flag. But people will choose.

We won’t guess a creator’s political opinions, religion, ethnic origin, health or sexual orientation. What creators choose to share about themselves is their call.

We won’t let an AI agent negotiate with a creator without saying it’s an AI, and who it works for.

We won’t let third parties train their models on our clients’ or creators’ data.

OUR AI COMMITMENTS

1. Human creators first. We will provide our users with tools that show how creators use AI, flag virtual influencers and estimate the share of AI-generated content (not just AI-edited) in a creator’s feed, so brands choose knowingly. Any message sent by AI through Kolsquare will say so, and on whose behalf.

2. Creators have a say about their data. Every creator will be able to see what Kolsquare shows about them, complete it and keep it up to date. When our brand safety tools flag something, they show the posts behind the flag. The brand or agency decides, and the creator can see what was flagged, in Kreators.

3. AI you can see, understand and switch off. Every AI output in Kolsquare will be labelled and come with the reasons behind it. Every user will have a log of the AI used on their account and the credits it consumed. We will be transparent about which models we use, and for what. And every account will be able to cap its use of generative AI, down to zero.

4. AI as light as possible. We will measure the footprint of the AI we use (energy, emissions, water) and show an estimate in the product, next to credit usage. It extends the campaign carbon calculator we already offer to our clients. We’ll use the smallest model that does the job, hosted close to our users when we can, and publish our method and our totals every year in our Mission Report. We’ll also take labour conditions in the data supply chain into account when we choose our AI providers.

5. Public action. With the European Influencer Marketing Alliance, and the local trade bodies as well as the rest of the industry, public authorities and regulators, we’ll push for clear AI disclosure in influence, public rules on how platforms and their algorithms decide what gets seen, stronger regulation of digital spaces, attention hygiene for children (at school and at home), and a more sober ecology of communication. And we’ll push back against a digital colonisation of the world by a handful of players. Every year, 1% of our revenue already goes to NGOs working on the limits of social media and influence. Part of it will go to this work.

WHAT THE AI LAW REQUIRES, AND WHERE KOLSQUARE GOES FURTHER

Since 2 August 2026, the EU AI Act requires that people know when they’re dealing with an AI, and that deepfakes be disclosed (Article 50). We will comply, like everyone should. But for us compliance is the floor. The five commitments above go further than the law asks, and we’ll report on them publicly, every year, including where we fall short.

WHAT WE DON’T KNOW YET

We don’t know how fast agents will change the way brands and creators work together. We don’t know how well AI-content detection will hold up at scale, so our estimates will always come with their margin of error. And we don’t yet know the real footprint of our own AI. That’s why we’re starting to measure it.

This is version 1. We’ll revise it every year with our team, creators, clients and partners. If you think we’ve missed something, or got something wrong, tell us.

Kolsquare. Entreprise à Mission since 2020, B Corp since 2024.

---

So here it is!

What it will probably cost us

I strongly believe that values only count when they cost something, so here’s my best guess at the bill. We’ll turn down some business, starting with AI twins and tools that promote virtual influencers, knowing a competitor or a new entrant will probably say yes. Some promises will be hard to keep: detecting AI-generated content, letting every account cap its use of generative AI, showing creators what our tools say about them. Our product and tech teams have reminded me of that several times, and I know they are right. And once we measure the footprint of our AI, we’ll have to live with what we find.

Lots of these features don’t exist yet. We’ll build them step by step, probably adjust how we deliver some of them along the way, and report every year in our Mission Report, including what we didn’t manage to do.

What I’d like influence to look like in a few years

A manifesto from one company won’t change an industry. But I can describe the one I’d like to work in.

A brand can see which part of a campaign was made with AI, and why its tools recommended one creator over another. A creator knows when they’re negotiating with a machine, and can check what platforms say about them. An agency’s juniors learn the craft before they learn the prompts. A teenager scrolling through a feed can tell what’s real. We all know, roughly, what our AI costs the planet. And because all of this is visible, people stop worrying about whether AI was involved. They just judge whether the content is good and honest.

None of this is out of reach. Here are a few ideas I think could get us there:

• brands that make a clear choice to favour authentic content and human creators, and say so;
• a common way to label AI content in influence, one that survives when a post moves from one platform to another;
• contract clauses that protect a creator’s face, voice and content by default;
• a simple rule for AI agents to say who they are and who they work for;
• shared ways to measure the footprint of AI in marketing.

The questions I can’t answer yet

• Will AI widen the creator economy, or concentrate it in fewer hands?
• When an AI answer quotes a creator, who should get paid?
• What should a creator always have the right to refuse?
• How do we keep trust when most content is at least partly synthetic?

I don’t think any one company can answer these alone. That’s why this is a version 1.

Quentin.