Source: Politico Europe
Introduction
European regulators face mounting challenges as generative artificial intelligence blurs traditional legal boundaries. Treating conversational tools as standard information retrieval systems offers only a partial solution for oversight bodies. When citizens interact with artificial intelligence for companionship or emotional support, existing regulatory frameworks reveal significant blind spots.
ChatGPT soumis à de nouvelles règles européennes aux multiples angles morts, raising urgent questions across the continent. While lawmakers attempt to rein in dominant technology platforms, the multifaceted nature of modern chatbots defies simple categorization. Experts warn that current oversight mechanisms fail to address the full spectrum of societal and psychological risks.
What Happened
The European Commission officially designated ChatGPT as a very large online search engine under the Digital Services Act. This classification subjects OpenAI to strict transparency and risk mitigation mandates comparable to established industry giants like Google and Bing. Companies failing to comply face severe financial penalties reaching up to six percent of their total annual global turnover.
However, this regulatory approach applies exclusively to the retrieval functions of the software. Conversational exchanges generated independently by the chatbot remain seemingly unaffected by this specific designation. Furthermore, authorities opted against classifying the technology as a very large online platform, a label traditionally reserved for social media networks and electronic commerce sites.
Background
Adopted in 2022, the Digital Services Act preceded the widespread commercial explosion of advanced conversational interfaces. Legal scholars describe the underlying technology as a hybrid entity combining search capabilities, platform features, and editorial traits. Traditional platform regulations rely on safe harbor principles exempting companies from liability for user-generated content. Conversations between humans and artificial intelligence challenge whether machine-generated dialogue fits existing legal definitions.
Simultaneously, the European Union oversees foundational artificial intelligence models through a separate legislative framework known as the AI Act. This parallel statute requires developers of general-purpose systems to evaluate and mitigate systemic risks. Authorities began enforcing these model-level security mandates following targeted inquiries directed at major artificial intelligence developers.
Timeline
| Date / Period | Regulatory Milestone |
|---|---|
| 2022 | Adoption of the Digital Services Act by European legislators. |
| August (Previous Year) | Implementation of risk evaluation mandates for general-purpose artificial intelligence models under the AI Act. |
| Late August | European Commission initiates enforcement actions by questioning artificial intelligence companies on safety protocols. |
| Monday | Executive branch classifies ChatGPT as a very large online search engine. |
Key Details
The narrow search engine classification limits the capacity of regulators to supervise risks associated with mental health, unlawful material, or election integrity. While querying candidates in a local election falls under regulatory scope, open-ended political discussions involving potential misinformation may escape oversight. European lawmakers spent nearly a year finalizing this designation due to classification difficulties.
Independent research indicates widespread reliance on conversational agents for personal health and counseling purposes. Insurance industry data reveals that sixty percent of adults globally utilize chatbots for therapeutic support. Concurrently, public scrutiny intensifies following tragic adolescent suicides allegedly linked to chatbot usage, prompting ongoing legal actions by grieving families.
Impact
Legislators emphasize that foundational risk assessments primarily target existential dangers rather than fundamental rights violations. Experts note that technical safeguards often overlook weapon development risks while neglecting psychological harms experienced by everyday users. Lawmakers underscore the necessity of complementary oversight to protect individuals developing dangerous emotional dependencies on companion chatbots.
Member of the European Parliament Christel Schaldemose highlighted specific dangers facing children, including addictive design elements and emotional reliance. Additional policy analysts argue that applying service regulations helps scrutinize software design choices rather than treating the application merely as an abstract model.
What Happens Next
European authorities intend to leverage complementary regulatory frameworks to supervise both underlying models and consumer applications. Officials confirm that regulatory scrutiny will extend from foundational technology to the concrete deployment of consumer services. Future developments will depend on the full enforcement of the Digital Services Act and the ongoing execution of the Artificial Intelligence Act.
