AI was the top opportunity in AVP’s 2026 DAM Trends Report. It was also the top risk.
23.7%
named AI as the opportunity for the year ahead
What it means: auto-tagging, predictive metadata, agentic AI on repetitive tasks, DAM as a content creation engine.
21.7%
named AI their top risk for the year ahead
What it means: AI pushed without clear use cases, layered onto unstable metadata and workflows, ahead of program maturity.
Excitement and caution about AI aren’t opposite instincts. They’re the same instinct, applied to two questions: what could this do for us, and are we ready to do it safely. Preparedness means answering both before AI outruns your foundations. If you want to unlock the power of AI, then you will need the foundation in place first.
Governance is one of the core foundations of AVP’s DAM Operational Model, a framework that blends program, product, and service management to guide organizations starting their DAM journey and to help established programs identify gaps. This piece focuses on governance because it is the foundation AI puts under the most pressure; to see how it works alongside the others, you can score your own program’s operational health.
If your organization is under pressure to move faster with AI, the highest-leverage move is to treat DAM fundamentals as an executive level capability, not an operational nice-to-have. With DAM fundamentals in place, you have:
- risk mitigation
- preparedness
- efficiency enablement and value realization
The goal is not to slow down AI … it’s to make AI safe and effective.
Governance
Governance is the mechanism for establishing ownership and clarifying the guidelines that make systems, assets, and data trustworthy, accessible, and valuable across the organization. It is what ensures DAM drives business value rather than business risk. AI has not changed the purpose of governance, but it has changed the conditions governance operates under, and that is why it has become more urgent.
The Rise of Shadow IT
Shadow IT is the use of technology (software, hardware, cloud services, apps, or devices) within an organization without the knowledge, approval, or oversight of the IT or security department. Unauthorized systems are an enterprise risk, threatening both cybersecurity and digital sovereignty. Shadow IT erodes control from the inside, often through browser-based tools and personal accounts that never show up in formal procurement or asset inventories. Regaining that control starts with discovery: using network telemetry, browser observability, SaaS metadata, and identity logs to reveal what’s actually in use across the organization.
In most cases, employees aren’t deliberately undermining policy; they’re simply trying to get work done faster. That’s exactly what makes shadow IT so dangerous. It usually enters through convenience, not malice. And once sensitive data leaves governed systems, the organization loses visibility into where it went, who accessed it, and whether it was retained or used to train AI models. The strongest response, therefore, is as much cultural as it is technical. Leaders need to balance employees’ expectations for tools that just work with the realities of running IT, and IT teams need the freedom to move quickly within governed guardrails so they can keep systems current and easy to use.
For DAM teams, shadow IT increasingly means shadow AI: assets drafted, edited, or tagged with unapproved tools that enter the DAM with no record of how they were made. Governance has to account for the content that arrives through side doors, not just the content that moves through approved workflows.
Governance – Why It’s More Urgent Now
DAM governance mechanisms such as approvals, rights, expiry, and relationships were built around the assumption that a human sits between the content and the customer as a final check. Even in a loosely governed system, a person’s judgment quietly absorbed a lot of the risk: “this looks outdated,” “let me confirm this is still cleared,” “this doesn’t fit this market.” That informal safeguard is what made imperfect governance tolerable for so long. AI removes it at the exact moment it increases speed and volume, and it does so in two ways.
First, output volume has outpaced oversight capacity. A team can review every asset behind a handful of campaigns, but it cannot review every AI-generated variant across hundreds of personalized, automated touchpoints. AI’s core value proposition, scale, is precisely what makes manual review unworkable as a backstop.
Second, AI is confidently wrong rather than hesitantly wrong. A person who is unsure about an asset’s status tends to pause or ask. An AI system working from incomplete or ambiguous governance data typically gives no signal of that uncertainty; it produces fluent, confident output whether or not the underlying content was valid. Ungoverned metadata does not sit passively in the system. It gets acted on as if it were fact.
The tools themselves are also changing the terms. In August 2026, Anthropic began embedding invisible watermarks in text generated by its Claude models, along with signed C2PA provenance metadata on image files, to meet its commitments under the EU AI Act. The marking applies worldwide, not only in the EU, and it persists through copying and pasting. In practice, this means content an employee drafts with an AI tool outside of approved channels may carry a machine-readable mark of AI involvement that the organization never recorded. Provenance is arriving whether or not your governance is ready for it, and the gap between what your DAM says about an asset and what the asset itself says becomes visible to anyone who checks.
The urgency, then, is not that AI is inherently riskier. It is that AI targets the exact failure mode most DAM programs have quietly relied on people to catch, and it does so at a speed and scale that turns latent, contained risk into fast, visible, multiplied risk. The response is to get your fundamentals in order, along with the people and processes that sustain them.
Governing Emerging AI Legislation
Trust sets expectations for positive future interactions and encourages participation with technology. Communicating the meaning and purpose of why a technology tool is being used will build trust with its audience and impact positive experiences. Trust in technology and the data flowing through all connected systems will lead to greater participation that will increase information’s value and utility. But artificial intelligence in our content, documentation, and marketing information is making all of this messier and more complicated, which raises a harder question: do we trust what we see and read?
From a legal point of view, there is some hope for the future as new legislation regarding AI creation and usage does take into account issues of “transparency” and “provenance,” as seen in the following important pieces of current 2026 legislation:
EU AI Act (2026) — Article 50
Providers of generative AI systems must mark AI-generated audio, image, video, and text in a machine-readable format so it can be detected as artificially generated, and organizations deploying AI must disclose deepfakes. The obligations apply to AI offered or used in the EU, regardless of where the company is headquartered.
California AI Transparency Act (2026)
Covered providers, meaning generative AI systems with more than one million monthly users in California, must embed a hidden (latent) disclosure in AI-generated image, video, and audio content, offer users a visible (manifest) disclosure, and provide a free detection tool. The law is enforced by the California Attorney General.
The standard both laws point to the C2PA (Coalition for Content Provenance and Authenticity)
Neither law mandates a single technical standard, but the C2PA (Coalition for Content Provenance and Authenticity) specification has become the most widely adopted way to meet these requirements. It embeds a cryptographically signed record of origin directly into the asset, making provenance tamper-evident and machine-readable throughout its lifecycle.
How to Protect Brand Integrity from Slop
If you want to see how a society is doing, take a look at its language. Just look at our current situation in our English vocabulary:
- Slop (n)
- digital content of low quality produced in quantity by AI
- Enshittification (n)
- the degradation in the quality and experience of online platforms over time, due to an increase in advertisements, costs, or features
AI slop (n) is really a governance failure wearing a technology label
Content that’s technically on-brand-adjacent (right colors, right product, plausible copy) but lacks the judgment, accuracy, or distinctiveness a human reviewer would have caught. Protecting against it isn’t about restricting AI use — it’s about making sure the same structural discipline discussed throughout this piece (rights, approval, relationships, context) gets applied to AI output with the same rigor as anything else entering the DAM, rather than treating AI-generated content as exempt because it’s “just a draft” or “just personalization.”
Here are three ways you can protect your brand integrity from AI slop:
-
Treat AI-generated content as a distinct asset class with its own governance.
The instinct is to let AI output skip the queue because “someone reviewed the prompt.” But prompting isn’t approval, and the AI removes exactly the review step that used to catch drift.
- AI-generated assets should carry explicit provenance metadata, such as the model used, the prompt or brief reference, and the generation date.
- Inherit “unapproved” by default until someone with actual authority signs off, the same as any other new asset.
-
Anchor generation to governed source assets, not open prompting.
Slop tends to come from AI generating “in the general direction of the brand” from a text description, rather than from approved brand assets, locked templates, and reference imagery. The fix is architectural:
- Use the DAM’s approved masters, brand guidelines, and locked style and voice references as the system of record.
- AI output should be linked back to the master assets and guidelines it was derived from, so drift is traceable and correctable at the source.
-
Distinguish AI-assisted content from AI-autonomous content.
Content where AI drafted and a human materially edited and approved is a different risk category from content an AI generated and distributed with no human touch. Treating these identically in your governance model is itself a source of slop — the low-oversight category should be the exception you deliberately choose for low-stakes use cases, not the default because it’s faster. Make brand and compliance rules machine-checkable, not just human-readable.
And don’t forget that AI governance is not a solo responsibility … it’s a healthy relationship between marketing, legal, IT, and the DAM team. In practice, the organizations that do this well set up a cross-functional governance group with a single accountable executive and give each function a defined slice of responsibility. If you need a place to start now, begin with an AI provenance and labeling policy: every asset or metadata field that AI touches is marked as such when it enters or changes in the DAM. Start there because it is the one step that becomes dramatically harder later, and because with AI, metadata is no longer a supporting detail. It is the governance.
Conclusion
The throughline: “using AI responsibly” isn’t a separate initiative from good DAM governance — it’s what good DAM governance looks like once AI is one of the things acting on your content. The organizations that avoid slop aren’t the ones being most cautious about using AI; they’re the ones whose rights, approval, relationship, metadata, and compliance layers were already rigorous enough that AI output has to pass through the same gates as everything else, rather than sliding in through a gap those disciplines never covered.
Frequently asked questions
What is DAM governance?
DAM governance is how an organization establishes ownership and sets the guidelines that make its systems, assets, and data trustworthy, accessible, and valuable. It covers mechanisms such as approvals, rights, expiry, and asset relationships.
What is shadow AI in Digital Asset Management?
Shadow AI is the use of unapproved AI tools to draft, edit, or tag assets. Those assets can enter the DAM with no record of how they were made, which creates rights, provenance, and brand risk.
Why does AI make DAM governance more urgent?
Most DAM programs relied on a person as the final check before content reached an audience. AI removes that check while increasing volume, and it acts on incomplete metadata with no signal of uncertainty.
What is C2PA, and why does it matter for a DAM?
C2PA is an open standard that embeds a cryptographically signed record of an asset’s origin. It has become the most widely adopted way to meet AI transparency laws, so a DAM needs to preserve this data through ingest, transformation, and export.
What is the first step toward AI governance in a DAM?
Start with an AI provenance and labeling policy, so every asset or metadata field AI touches is marked when it enters or changes in the DAM. It is the step that becomes hardest to add later.