{"id":123040,"date":"2025-02-11T20:18:14","date_gmt":"2025-02-11T19:18:14","guid":{"rendered":"https:\/\/ceadigilaw.org\/blog\/the-withdrawal-of-the-liability-directive-in-ai-a-strategic-turn-or-an-eu\/"},"modified":"2025-02-11T20:18:14","modified_gmt":"2025-02-11T19:18:14","slug":"the-withdrawal-of-the-liability-directive-in-ai-a-strategic-turn-or-an-eu","status":"publish","type":"post","link":"https:\/\/ceadigilaw.org\/ar\/blog\/the-withdrawal-of-the-liability-directive-in-ai-a-strategic-turn-or-an-eu\/","title":{"rendered":"The withdrawal of the AI Liability Directive: a strategic shift or a regulatory step backwards for the EU?"},"content":{"rendered":"
\n
\n
By <\/span>Antonio Tejeda Encinas<\/span><\/span><\/a><\/span>, President of the <\/span>Euro-American Committee of Digital Law - CEA Digital Law<\/span><\/span><\/a><\/span> and CEO of <\/span>Meta Channel Corp<\/span><\/span><\/a><\/span><\/em><\/div>\n<\/div>\n
\n
<\/div>\n

Introduction: the European paradox in AI regulation<\/h3>\n<\/div>\n
\n
<\/div>\n
Europe has positioned itself as the epicentre of artificial intelligence (AI) regulation. With the recent adoption of the EU Artificial Intelligence Act (Regulation (EU) 2024\/1689), or AI Act, the European Union (EU) has made clear its commitment to safety and ethics in the development of these technologies. However, the unexpected withdrawal of the AI Liability Directive<\/strong> by the European Commission on 11 February 2025 has created a profound contradiction: how can the EU protect citizens through an extremely strict regulation while removing a key piece of its legal framework, one that was meant to guarantee redress for harm caused by AI?<\/div>\n<\/div>\n
\n
<\/div>\n
This decision is not merely a regulatory adjustment. It is a move on the geopolitical chessboard of AI, where China and the United States are forging ahead with more flexible rules and massive investment. Europe, meanwhile, stands at a crossroads: should it bet on innovation at the expense of consumer protection, or reinforce its ethical principles at the risk of falling behind technologically?<\/div>\n<\/div>\n
\n
In this analysis, we unpack what this withdrawal means, its practical consequences and the dilemma facing the EU in the global AI race.<\/div>\n<\/div>\n
\n
<\/div>\n

1. The withdrawal of the directive: what happened and why?<\/h3>\n<\/div>\n
\n
<\/div>\n
On 11 February 2025, the European Commission announced that it was withdrawing the proposed AI Liability Directive, which had been put forward in 2022 to make it easier to claim compensation for harm caused by AI. The proposal provided for a presumption of causality, which would have made it easier for victims to prove that a defective AI system had caused them harm.<\/div>\n<\/div>\n
\n
<\/div>\n

The official reasons<\/h4>\n<\/div>\n
\n
<\/div>\n
The Commission justified the withdrawal on three main grounds:<\/div>\n<\/div>\n
\n
<\/div>\n
1. A lack of consensus among Member States on the need for a specific AI directive.<\/div>\n<\/div>\n
\n
2. An excessive regulatory burden, given that the AI Act already covered key safety aspects.<\/div>\n<\/div>\n
\n
3. Concerns about European competitiveness, particularly following criticism from the tech industry and pressure from international partners such as the US.<\/div>\n<\/div>\n
\n
<\/div>\n

The real reasons: caught between industry and geopolitics<\/h4>\n<\/div>\n
\n
<\/div>\n
The official arguments may seem reasonable, but the truth is that big tech lobbied hard against this legislation. Companies such as Google, Microsoft and Meta have repeatedly warned of the risks of an \u2018overly restrictive\u2019 regulatory environment in Europe. At the same time, the US and China have stepped up their investment in AI under far more flexible rules, which could leave Europe at a competitive disadvantage.<\/div>\n<\/div>\n
\n
<\/div>\n

2. Practical consequences of the withdrawal: who wins and who loses?<\/h3>\n<\/div>\n
\n
<\/div>\n
Scrapping this directive has legal, economic and strategic implications. Let us look at its effects from a practical perspective, with examples.<\/div>\n<\/div>\n
\n
<\/div>\n
A. For consumers: left unprotected and without legal recourse<\/strong><\/div>\n<\/div>\n
\n
<\/div>\n
One of the main problems with withdrawing the directive is that victims of AI-related harm will find it much harder to obtain compensation. Take a straightforward example:<\/div>\n<\/div>\n
\n
<\/div>\n
A medical AI system wrongly diagnoses a patient with cancer and recommends unnecessary surgery.<\/div>\n<\/div>\n
\n
<\/div>\n
Without a specific liability directive, the burden of proof falls entirely on the injured party.<\/div>\n<\/div>\n
\n
<\/div>\n
Proving that the AI failed, that there was negligence and that this caused harm may be almost impossible without access to the system\u2019s data.<\/div>\n<\/div>\n
\n
<\/div>\n
This legal vacuum benefits tech companies and harms users, who will have to rely on a patchwork of national rules.<\/div>\n<\/div>\n
\n
<\/div>\n
B. For AI companies: regulatory relief, but not without risks<\/strong><\/div>\n<\/div>\n
\n
<\/div>\n
For European start-ups and tech companies, the withdrawal of this directive offers breathing space. A less restrictive regulatory framework could:<\/div>\n<\/div>\n
\n
<\/div>\n
Attract more investment into Europe\u2019s AI sector.<\/div>\n<\/div>\n
\n
<\/div>\n
Avoid mass litigation against AI developers.<\/div>\n<\/div>\n
\n
<\/div>\n
Make it easier to bring AI products to market without fear of constant lawsuits.<\/div>\n<\/div>\n
\n
<\/div>\n
This relief may be short-lived, however. Without a harmonised directive, each EU country will regulate differently, which may complicate cross-border trade and create legal uncertainty.<\/div>\n<\/div>\n
\n
<\/div>\n
C. For the EU: at odds with its own regulation<\/strong><\/div>\n<\/div>\n
\n
<\/div>\n
Herein lies the great contradiction: the EU has tightened the regulatory framework for AI with the AI Act, yet it has removed a key protection mechanism. In other words:<\/div>\n<\/div>\n
\n
<\/div>\n
On the one hand, it imposes tough restrictions on AI developers to make sure their products are safe.<\/div>\n<\/div>\n
\n
<\/div>\n
On the other, it scraps the rule that determined who is liable when AI fails.<\/div>\n<\/div>\n
\n
<\/div>\n
This double standard creates uncertainty and may weaken the EU\u2019s standing as a global regulatory benchmark.<\/div>\n<\/div>\n
\n
<\/div>\n

3. The global race: Europe versus the United States and China<\/h3>\n<\/div>\n
\n
<\/div>\n
The EU\u2019s decision cannot be understood without looking at the international chessboard of AI.<\/div>\n<\/div>\n
\n
<\/div>\n
China has pursued a model of technological self-sufficiency, investing billions in AI and avoiding any regulation that might slow its development.<\/div>\n<\/div>\n
\n
<\/div>\n
The US, despite ethical concerns, has opted for a minimal-regulation approach, allowing tech giants such as OpenAI and Google to dominate the market.<\/div>\n<\/div>\n
\n
<\/div>\n
Europe, by contrast, has sought to lead on AI regulation, but this withdrawal could cost it influence in setting global standards.<\/div>\n<\/div>\n
\n
<\/div>\n
A practical example: the DeepSeek case<\/strong><\/div>\n<\/div>\n
\n
<\/div>\n
The Chinese AI model DeepSeek recently outperformed Western models on efficiency and cost. How did it manage this? With a flexible regulatory framework and access to vast amounts of data. Europe, with its strong emphasis on data protection and security, runs the risk that its companies will be unable to compete on a level playing field.<\/div>\n<\/div>\n
\n
<\/div>\n

4. A critical perspective: \u2018good cop\u2019 vs \u2018bad cop\u2019<\/h3>\n<\/div>\n
\n
<\/div>\n
To assess this decision in a balanced way, let us consider two opposing perspectives.<\/div>\n<\/div>\n
\n
<\/div>\n
Good cop: an opportunity to rethink regulation<\/strong><\/div>\n<\/div>\n
\n
<\/div>\n
A lighter regulatory burden for AI start-ups.<\/div>\n<\/div>\n
\n
Greater flexibility to adapt the rules in future.<\/div>\n<\/div>\n
\n
It keeps Europe from falling behind in the AI race.<\/div>\n<\/div>\n
\n
From this point of view, the EU has taken a pragmatic decision, adjusting to market realities.<\/div>\n<\/div>\n
\n
<\/div>\n
Bad cop: a step backwards for consumer protection<\/strong><\/div>\n<\/div>\n
\n
<\/div>\n
Citizens left wholly unprotected against AI-related harm.<\/div>\n<\/div>\n
\n
Legal fragmentation across Member States.<\/div>\n<\/div>\n
\n
A loss of credibility for the EU as a regulatory leader in AI.<\/div>\n<\/div>\n
\n
Seen in this light, withdrawing the directive is a strategic mistake that leaves a legal vacuum that is hard to justify.<\/div>\n<\/div>\n
\n
<\/div>\n

5. What should the EU do now?<\/h3>\n<\/div>\n
\n
<\/div>\n
The European Union faces a complex challenge. It cannot afford a regulatory environment that discourages innovation, but nor can it ignore the need to protect citizens from the risks of AI.<\/div>\n<\/div>\n
\n
<\/div>\n
To strike a balance between the two, the EU should consider:<\/div>\n<\/div>\n
\n
<\/div>\n
1. Creating an integrated liability framework within the AI Act.<\/div>\n<\/div>\n
\n
2. Encouraging transparency in AI systems, making it easier to access evidence in litigation.<\/div>\n<\/div>\n
\n
3. Committing to a competitiveness strategy, increasing investment in AI without compromising its regulatory model.<\/div>\n<\/div>\n
\n
<\/div>\n
Withdrawing this directive is a risky gamble. It can be read as a strategy to strengthen Europe\u2019s AI industry or as a capitulation to market pressure. Time will tell whether it was a brilliant tactical move or a serious regulatory blunder.<\/div>\n<\/div>\n
\n
<\/div>\n
What is clear is that the debate on AI liability is far from over, and the EU will have to rethink its strategy if it does not want to end up trapped between over-regulation and technological irrelevance.<\/div>\n<\/div>\n<\/div>\n","protected":false},"excerpt":{"rendered":"

By Antonio Tejeda Encinas, President of the Euro-American Committee of Digital Law - CEA Digital Law and CEO of Meta Channel Corp Introduction: the European paradox in AI regulation Europe has positioned itself as the epicentre of artificial intelligence (AI) regulation. With the recent adoption of the EU Artificial Intelligence Act (Regulation (EU) 2024\/1689), or […]<\/p>\n","protected":false},"author":2,"featured_media":113147,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"wds_primary_category":0,"wds_primary_cea_women":0,"footnotes":""},"categories":[1126],"tags":[],"cea_women":[],"class_list":["post-123040","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-inteligencia-artificial-ar"],"acf":[],"_links":{"self":[{"href":"https:\/\/ceadigilaw.org\/ar\/wp-json\/wp\/v2\/posts\/123040","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/ceadigilaw.org\/ar\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/ceadigilaw.org\/ar\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/ceadigilaw.org\/ar\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/ceadigilaw.org\/ar\/wp-json\/wp\/v2\/comments?post=123040"}],"version-history":[{"count":0,"href":"https:\/\/ceadigilaw.org\/ar\/wp-json\/wp\/v2\/posts\/123040\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/ceadigilaw.org\/ar\/wp-json\/wp\/v2\/media\/113147"}],"wp:attachment":[{"href":"https:\/\/ceadigilaw.org\/ar\/wp-json\/wp\/v2\/media?parent=123040"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/ceadigilaw.org\/ar\/wp-json\/wp\/v2\/categories?post=123040"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/ceadigilaw.org\/ar\/wp-json\/wp\/v2\/tags?post=123040"},{"taxonomy":"cea_women","embeddable":true,"href":"https:\/\/ceadigilaw.org\/ar\/wp-json\/wp\/v2\/cea_women?post=123040"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}