Investors once began their journey with a platform. Increasingly, they begin with a question and that question is asked to a machine. Across insurance markets, AI systems have already crossed the line from information to intermediation. Crowdfunding may be next.
When Europe adopted the European Crowdfunding Service Providers Regulation (ECSPR), it set out to build a cross‑border digital investment market that could protect retail investors without stifling innovation. The regulation is unusually sophisticated: it assumes investors discover opportunities online, anticipates cross‑border flows, and requires standardised key investment information sheets, appropriateness assessments, risk warnings and governance standards designed for a digital marketplace. What it did not anticipate is that the investor’s first interaction may no longer be with a crowdfunding platform at all.
Today, the journey often begins elsewhere.
- “What is the best European crowdfunding platform?”
- “Which renewable energy projects have the highest expected returns?”
- “I have €5,000 to invest. Should I use crowdfunding or ETFs?”
- “What is the safest real estate crowdfunding investment?”
Only a few years ago, these questions would have led investors to a search engine, a financial newspaper or perhaps an adviser. Increasingly, they are addressed to artificial intelligence systems. Large language models have become conversational interfaces that filter, compare, rank and recommend. This shift deserves attention. Not because artificial intelligence is inherently problematic, but because it quietly changes where investment decisions are actually being shaped.
Lessons from the Insurance Market
Earlier this summer, our colleagues at FECIF (EUROCROWD is a member of FECIF) raised this point in the context of insurance distribution. They documented cases where AI systems recommended named insurance products, complete with tariff designations, suitability reasoning and directions on where contracts could be concluded, i.e. conduct that normally requires authorisation under the Insurance Distribution Directive. Their concern was simple: if equivalent functions are performed, should equivalent rules apply?
Crowdfunding faces a remarkably similar question, just under a different legal framework. The ECSPR regulates what happens once an investor reaches a platform. Before investing, non‑sophisticated investors receive risk warnings, undergo an appropriateness assessment, and are invited to reflect on their ability to bear losses through simulations and, where applicable, a reflection period. Project information must be standardised. Marketing must be fair, clear and not misleading. Platforms are supervised. Complaints procedures exist. Governance obligations are extensive.
These protections represent years of legislative work and one of Europe’s most comprehensive investor protection regimes for alternative finance. Yet none of these safeguards apply before the investor arrives. If an AI system has already narrowed the market from hundreds of platforms to three, highlighted one project as particularly attractive, explained why expected returns appear favourable, and reassured the user that the investment matches their stated objectives, much of the practical investment decision may already have been made.
From Destination to Journey
The ECSPR governs the destination. Artificial intelligence increasingly governs the journey. And that shift is not trivial. European financial regulation has long recognised that influence over investment decisions matters. The distinction between factual information and regulated investment advice has never been accidental. Under MiFID and related legislation, considerable effort has gone into defining where information ends and personal recommendation begins. Those boundaries are neither simple nor static.
Artificial intelligence now tests them in new ways. A chatbot does not present itself as an adviser. It simply answers questions. Yet those answers can be highly persuasive. They synthesise vast amounts of information, present conclusions with confidence and often remove the uncertainty that might otherwise lead investors to seek professional advice.
Behaviourally, that influence can be as strong as traditional financial intermediation. Legally, the position is far less clear. One should resist the temptation to conclude too quickly that AI systems are providing regulated investment advice. The legal threshold is specific and deserves careful analysis. Many interactions will remain general information. Others may not. The difficult cases increasingly sit somewhere in between. But that ambiguity alone should interest European supervisors.
When Analysis is replaced by Machine Learning
There is another dimension worth attention. Artificial intelligence does not discover investment opportunities through objective market analysis. It learns from digital information. Providers with stronger online visibility, better structured data, clearer documentation and more effective Generative Engine Optimisation (GEO) naturally appear more often in AI‑generated responses. FECIF’s own testing found that identical queries submitted to the same systems produced inconsistent recommendations across sessions, underscoring that optimisation, not objectivity, often drives visibility. This creates a competitive dynamic almost entirely outside existing regulatory assumptions. Success may increasingly depend not only on the quality of projects or investor outcomes, but on how effectively platforms communicate with machines rather than humans.
Investors will not appreciate this distinction. To them, an AI‑generated recommendation appears authoritative to some degree. Few will understand why Platform A appears repeatedly while Platform B rarely does, or whether this reflects quality, digital optimisation or simply the statistical behaviour of large language models. Transparency becomes difficult precisely because the recommendation process is opaque. The implications extend beyond competition.
Imagine an investor asking which European crowdfunding opportunity best matches a cautious profile. An AI recommends a specific project. The investor proceeds and suffers substantial losses. Where does accountability reside? Not with the AI provider under the ECSPR. Not necessarily with the crowdfunding platform if the recommendation originated elsewhere. Certainly not with the regulatory mechanisms designed around interactions occurring within the platform itself. As FECIF’s supervisory inquiries showed, every authority recognised the issue; none considered itself the competent body to act. The result is a supervision gap precisely at the point where investor influence now occurs.
Acting with Care to embrace the Unavoidable
Europe has invested considerable effort in constructing clear lines of responsibility within crowdfunding markets. Artificial intelligence introduces the possibility that those lines become blurred before the regulatory framework even begins to operate. But lets be careful. This is not an argument against artificial intelligence. AI will almost certainly become indispensable across financial services in one way or another. Crowdfunding platforms are already exploring AI‑assisted due diligence, customer support, fraud detection, document analysis and investor education. Used responsibly, these technologies seem to promise substantial gains in efficiency and accessibility.
This is also not an argument for extending licensing requirements to every conversational AI. Rather, we think it is an invitation to examine whether Europe’s regulatory efforts still corresponds with the realities of digital investor behaviour. Financial regulation has historically followed functions rather than technologies. When new technologies perform economically equivalent roles, legislators eventually ask whether existing rules remain appropriate. That discussion accompanied online brokerage, algorithmic trading, robo‑advice and crypto‑assets. Artificial intelligence is simply the latest iteration of the same underlying question, and surely not the last.
Ensuring Investor Protection reaches Investors
The ECSPR remains, in our view, an exceptionally well‑designed regulation. Yet like any legislative framework, it reflects the assumptions of its time (one decade ago). One of those assumptions was that the regulated crowdfunding platform would be the principal gateway through which investors discovered opportunities. That assumption is becoming less certain. The investor increasingly meets the chatbot long before the platform. If that chatbot materially shapes how investors discover, compare and select crowdfunding opportunities, then Europe faces a question that extends well beyond crowdfunding itself. Consumer protection is not determined solely by what happens once investors enter a regulated environment. It also depends on who shapes their decisions before they arrive.
That is a question worth asking today, while artificial intelligence is still becoming embedded in financial decision‑making, not after new habits have become impossible to reverse. For the crowdfunding sector, the issue is less whether AI belongs in European finance. The more important question is whether the investor protections embedded within the ECSPR still reach investors at the point where their decisions are increasingly being formed. If the answer is becoming less certain, then perhaps the next evolution of Europe’s crowdfunding rulebook has already announced itself via chatbot.
