Beyond AI Adoption: The Control Questions Now Shaping Intellectual Property Strategy

Businesses are moving artificial intelligence from experiments into core workflows that affect product development, branding, content creation, research, enforcement, and legal services. Much of the business conversation around AI has focused on productivity, pricing models, and efficiency. Those considerations still matter, but they no longer capture the full risk profile. As AI becomes embedded in business operations, companies are more deeply confronting questions that reach ownership, evidence, vendor risk, data exposure, governance, and cross-border enforcement.

Those issues are increasingly converging around a broader question of control: how AI is governed inside organizations, how it is regulated by governments and institutions, and who controls the data, models, infrastructure, standards, and deployment channels that make AI possible.

AI governance, AI policy, and AI sovereignty are related, but they are not interchangeable.

AI governance concerns how AI systems are directed, monitored, managed, and reviewed. In practical terms, governance asks how an organization decides which AI tools may be used, who approves them, what controls apply, how human review remains meaningful, what risks are tracked, and what happens when something goes wrong.

AI policy concerns the substantive legal, regulatory, institutional, and strategic choices that shape AI. Policy asks what rules should apply, what objectives should be served, what risks should be prioritized, and what positions governments, courts, regulators, companies, and professional organizations should take.

AI sovereignty is not a synonym for AI control. It is a narrower but increasingly important expression of control: the question of who has meaningful authority over the data, compute, models, infrastructure, standards, procurement, security, and deployment channels that make AI possible.

For governments, AI sovereignty may involve national strategies around infrastructure, localization, security, standards, innovation, cultural identity, and technological independence. For businesses, the same concept becomes operational: where data goes, who operates the model, what infrastructure supports the system, whether activity can be audited, whether vendors or environments can be changed, and whether AI workflows comply in each jurisdiction where the business operates.

For IP owners, those distinctions matter because control over AI tools, data, evidence, and decision points may determine whether rights can be owned, proven, protected, and enforced.

AI is changing how intellectual property rights are created and how IP assets are protected, challenged, and enforced. Brands may be misused in synthetic advertisements, public figures may be impersonated through deepfake endorsements, and confidential or copyrighted materials may surface in AI-assisted workflows, training sets, or outputs. Trademark offices and enforcement authorities are also beginning to evaluate or use AI for examination, classification, image search, fraud detection, and other functions.

The control dimension becomes concrete when those issues cross borders. Consider a company using an AI vendor to generate advertising assets for a global campaign. The prompts may include product plans, brand concepts, or confidential positioning. The output may include synthetic imagery. The system may store logs in one jurisdiction, rely on model infrastructure in another, and produce content used in markets with different rules for disclosure, labeling, evidence, and enforcement. If a dispute later arises, the company may need to reconstruct what was created, when it was created, what data was used, who reviewed it, whether confidential information was exposed, and whether the same use is treated differently in different jurisdictions.

That is not merely an AI-use question. It is a provenance, evidence, and control question.

Patent law is being forced to confront AI’s role in innovation. AI-assisted invention raises difficult questions about inventorship, ownership, disclosure, and whether existing patent frameworks adequately account for inventions developed with significant machine assistance. Those questions affect how companies document research and development, evaluate patentability, allocate ownership, and preserve value in innovation pipelines.

Technological infrastructure questions are being raised as well. If autonomous AI agents begin operating at Internet scale – placing orders, negotiating transactions, scraping content, initiating takedowns, or interacting with platforms on behalf of companies or consumers – rights holders may eventually need reliable ways to identify who or what initiated the activity, authenticate the agent’s authority, trace the system or account behind it, and preserve evidence across a variety of systems. That is a technology implementation issue, but it is also a legal, accountability, and risk-management issue.

Governments are examining whether reliance on AI systems developed and trained in a limited number of markets could gradually flatten local language, culture, and national identity. For IP owners, that makes sovereignty more than a question of infrastructure or regulation; it also asks whose cultural references, creative norms, and consumer expectations are being embedded in the tools that shape global commerce.

I recently began serving on the International Trademark Association’s (INTA) Technology Transformation Committee, including its Trends in Technology Subcommittee.

The Technology Transformation Committee monitors evolving technologies, such as artificial intelligence, NFTs, blockchain, Web3, and other developments, and assesses their impact on intellectual property rights. The committee also helps develop balanced, coordinated, and effective education, policies, and advocacy strategies relating to those technologies.

Within that broader effort, I am part of the Artificial Intelligence Working Group, which is examining several substantive areas, including responsible AI governance and safe use in IP practice; AI-driven infringement and enforcement; trademark and IP office modernization; transparency, provenance, and disclosure; and AI policy, sovereignty, and global legal trends. The last of those areas, in which I am participating, examines how evolving approaches to AI infrastructure, data governance, compute capacity, technical standards, localization, national security, innovation policy, regulation, and intellectual property may affect IP systems across jurisdictions.

This work is not about assuming that one policy answer will fit every technology, industry, or market. It is about understanding where legal regimes may converge, where they may diverge, and where rights holders may need clearer guidance before rules harden in ways that are expensive to unwind.

These issues cannot be resolved through legal analysis alone. Many of the legal consequences of AI turn on how AI systems are built, deployed, operated, and controlled.

Before entering the legal field, I worked as an electrical engineer on cellular communications and surveillance systems, where I used early machine-learning and security techniques well before “AI” became part of the daily business vocabulary. That work required thinking carefully about predictive systems, secure data transmission, failure points, and how technical events could be identified, preserved, and later understood. That background continues to shape how I counsel clients on AI, intellectual property, and emerging technology issues.

In addition to advising on whether a client can exert ownership over works created with AI tools, a lawyer evaluating AI-related risk should be asking technical and operational questions: What does the tool actually do? What data does it ingest? Where does that data go? Are prompts, uploads, outputs, or user interactions logged? Can those logs be audited later? What model or vendor touched the information? What human review occurred? What contractual obligations apply? What controls exist if the output is inaccurate, infringing, confidential, or challenged?

Those questions matter because AI risk often hides in implementation details. A company using AI for research and development, product design, content generation, brand monitoring, software development, or enforcement needs a practical framework that accounts for ownership, confidentiality, vendor diligence, risk allocation, documentation, governance, and the commercial value of the assets being created or protected.

Responsible AI adoption requires more than enthusiasm for new tools. It requires aligning the technology, the business objective, the legal risk, and the governance framework.

The legal profession is going through the same transformation our clients are experiencing. Law firms are evaluating how AI can improve workflows, enhance research, surface institutional knowledge, streamline drafting, support project management, and create more value-driven service models.

That practical experience matters for client counseling. A form policy or template can be useful, but it is only a starting point. Effective AI advice depends on understanding how tools are selected, deployed, governed, restricted, monitored, and reviewed in real workflows.

Fennemore’s work through Project BlueWave AI and DOT (Dynamic Optimization Technology), respectively, the firm’s AI initiative and secure internal AI assistant, gives us practical experience with many of the same implementation questions clients are now confronting: how to evaluate tools, preserve confidentiality, maintain human review, document decisions, allocate vendor risk, and build governance structures that can withstand scrutiny.

Not every organization should use the same tools or adopt AI in the same way. Meaningful AI counseling requires a working understanding of both the technology and the operational environment in which it is used.

I recently attended INTA’s Annual Meeting, where AI was woven into conversations about brand strategy, enforcement, rights protection, legal practice, technology infrastructure, patentability, and commerce. Much of the discussion focused on AI ownership, tools, and use cases. But the hidden gems were often about control: who controls the data, the evidence, the infrastructure, and the systems that businesses will increasingly rely on to create and protect value.

Many businesses that were initially cautious have begun adopting AI because it is increasingly embedded in tools and platforms they already use. But incorporation is not the same as control. In many organizations, the hardest questions are only beginning to surface.

The organizations best positioned for what comes next will be those treating AI as an enterprise capability, not a series of disconnected experiments. They will ask early how assets are being used, how outputs are documented, who owns the resulting work, what vendor obligations apply, which jurisdictions matter, where enforcement risks exist, and what governance structure is needed to make the workflow defensible later.

The next phase of IP strategy will depend on whether companies can document, govern, and defend AI-assisted creation and enforcement, including across jurisdictions. That is where law, technology, and business strategy now meet—and where organizations that pair innovation with discipline will be best positioned to succeed.

The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.