The year 2024 has reshuffled the cards in the global technological landscape. Between the entry into force of the first binding regulatory framework on artificial intelligence in Europe and the maturation of technologies that have long remained experimental, the lines have shifted in areas that the usual year-start assessments had not anticipated.
European AI Act: the regulatory framework that changes the game for AI innovation
The Regulation (EU) 2024/1689 on artificial intelligence was published in the Official Journal on July 12, 2024, and came into effect on August 1, 2024.
This text is not a statement of intent. It introduces a classification of AI systems by risk level, with direct consequences on the use cases permitted in Europe. Practices deemed to have “unacceptable risk” (social scoring, certain forms of behavioral manipulation, real-time biometric surveillance without safeguards) will become legally prohibited starting in 2025.
For businesses, the ramp-up is gradual until 2028. Transparency, documentation, and governance obligations for general-purpose AI models and high-risk systems will start to apply between 2025 and 2026.
The penalties foreseen can reach up to 35 million euros or 7% of global turnover. This level of penalty exceeds that of the GDPR and signals a political will to influence the technical choices of developers, even outside Europe.
Several use cases that “trend” articles still presented as promising thus find themselves in a gray area, or even closed off in the European market. Field feedback diverges on this point: some companies anticipate compliance, while others adopt a wait-and-see approach pending detailed guidelines. Following the news on these topics at tech articles on Blog Too allows for measuring the extent of these adjustments over the months.
Generative AI in business: beyond the hype
Generative AI has dominated technological conversations in 2023 and then in 2024. Capgemini, in its TechnoVision report, identified this technology as the main driver of transformation for organizations. The reality of adoption is more nuanced.
Language models have indeed improved. Their integration into business workflows (customer service, technical documentation writing, data analysis) is progressing in large companies. However, the costs of infrastructure and fine-tuning remain a barrier for medium-sized organizations. The actual effectiveness of these tools heavily depends on the quality of training data and the ability of teams to formulate relevant queries.
Small businesses can leverage generative AI provided they identify targeted use cases rather than aiming for a comprehensive transformation. A marketing content generation tool or accounting automation does not require the same resources as a banking fraud detection system.
Known limitations of generative models
Hallucinations (factually incorrect responses but confidently stated) remain a structural problem. The available data do not allow for concluding that this flaw will be resolved in the short term. Companies deploying these tools in production systematically add layers of human verification, which reduces the promised productivity gains.
Quantum computing and blockchain: two distinct trajectories
Quantum computing and blockchain have been featured in all technological overviews for several years. Their trajectory in 2024 deserves separate examination, as they are not at the same stage of maturity.
Quantum computing is gradually emerging from the laboratory, but commercial applications remain limited to niches (molecular simulation, logistical optimization for very large datasets). Current machines do not yet have the stability required for large-scale calculations under real-world conditions.
Blockchain, on the other hand, has a mature technological foundation. Industrial applications exist (supply chain traceability, document certification, digital rights management). Public adoption remains low, hindered by usability complexity and the persistent association with speculation on cryptocurrencies.
- In quantum computing, investments are focused on error correction and qubit stabilization, without a reliable timeline for scaling up.
- In blockchain, industrial cloud platforms now integrate ready-to-use modules, reducing the technical barrier for businesses.
- In both cases, development heavily depends on national regulatory choices, particularly regarding data security and digital sovereignty.
Transparent screens and robotics: consumer innovations scaling up
CES 2024 highlighted two categories of products that are moving beyond the prototype stage. Transparent screens, notably championed by Korean manufacturers, have reached a level of brightness and contrast that makes them usable in commercial contexts (storefronts, in-store displays, urban furniture). This is no longer a mere curiosity.
Consumer robotics has reached a similar milestone. The robot vacuums showcased at CES now incorporate vision-based navigation systems and real-time adaptation capabilities to the environment. The convergence between embedded AI and miniaturized sensors enables more reliable autonomous behaviors than previous generations.
Sustainable technologies and energy efficiency
A less publicized but structuring axis concerns sustainable technology. Industrial cloud platforms are evolving towards more precise management of energy consumption. Several providers offer carbon footprint measurement tools directly integrated into development environments.
This trend responds to regulatory pressure (mandatory non-financial reporting for large European companies) as much as to demand from end customers. Sustainability is becoming a criterion for technological selection, not just a marketing argument.
The year 2024 stands out for this overlay of technical maturations and new regulatory constraints. The framework established by the AI Act, the still tangible limitations of generative AI in production, and the distinct trajectories of quantum and blockchain create a landscape where each technological choice also entails a precise regulatory and budgetary analysis.
