I focus here on artificial intelligence, an area that continues to generate intense debate and a degree of uncertainty regarding the future of legal professions.
The legaltech ecosystem experienced a decisive turning point following the ChatGPT phenomenon in November 2022. Although natural language processing technologies were already used by certain large consulting firms and legal service providers, their adoption remained limited. At that time, artificial intelligence applied to law was primarily associated with predictive justice.
From 2022 onwards, lawyers and clients alike gained access to generative AI tools capable of significantly accelerating a wide range of legal tasks. The first versions of now well-known models and platforms (such as ChatGPT, Mistral or Claude) were still imperfect, but their impact on professional practices was immediate. For the first time, a broadly accessible technology demonstrated its ability to assist with drafting, summarisation, research and analysis in a way that was directly usable by legal professionals.
The most agile lawyers quickly sought to understand how these technologies could be integrated into their daily work. Until mid-2025, this phase was largely characterised by widespread experimentation within a fragmented legaltech landscape. Numerous generative AI solutions emerged simultaneously, with uneven levels of maturity, security and legal relevance. Law firms often tested multiple tools in parallel, without yet having the perspective necessary to assess their long-term value or to integrate them coherently into existing workflows.
By the end of 2025, this landscape has significantly narrowed and matured. Usage has progressively shifted from exploratory testing to a smaller number of identified and operational use cases. At the same time, established legal publishers and historical market players have integrated generative AI into their platforms, embedding it within trusted databases and familiar research environments. This evolution has contributed to a re-centralisation of practices around more reliable and professionally aligned solutions, marking a transition from dispersed experimentation to structured adoption.
A consensus has gradually emerged among practitioners who have invested time in understanding these tools: generative AI does not signal the end of the legal profession, but rather challenges those who refuse to engage with it. Just as it would be difficult today to rely on a lawyer unwilling or unable to use online legal databases or search engines, it is becoming increasingly difficult to justify a complete rejection of AI-assisted tools. The real differentiating factor is no longer access to technology, but the ability to use it critically, securely and in a manner consistent with professional standards.
The integration of generative AI into legal practice represents a genuine opportunity for lawyers to refocus on analysis, strategy and client support. This opportunity nevertheless comes with risks that must be anticipated and managed.
- Security remains a central concern. While the use of free and public tools continues to raise confidentiality issues, experience has shown that the main risks now stem from the multiplication of poorly governed solutions, the use of personal accounts for professional purposes and insufficiently controlled integrations. As a result, AI adoption has become a matter of governance, internal policy and awareness, rather than a purely technical issue.
- Reliability also remains a key issue, but its nature has evolved. In 2025, the frequency of hallucinations has decreased significantly in well-defined legal use cases, particularly where generative AI is combined with trusted sources, internal databases and structured prompting. The main risk is no longer the production of obviously incorrect outputs, but rather of answers that appear plausible while being incomplete or subtly inaccurate. This reinforces the necessity of systematic human oversight and professional judgment.
- The impact of generative AI on training and career development within law firms must be assessed with caution. In practice, no legal task has been fully automated, in particular due to the need for continuous verification, contextualisation and professional responsibility. While generative AI can accelerate or assist certain activities traditionally performed by junior lawyers, it does not remove the need for human analysis and judgment. These developments nonetheless require anticipation, as the nature of learning evolves alongside the tools used. Training must increasingly focus on the ability to frame legal issues, critically assess AI-generated outputs and ensure their accuracy and relevance, rather than on the mechanical execution of tasks. The challenge therefore lies not in a loss of learning opportunities, but in adapting training models to reflect these changing practices.
- Inequality in access to and familiarity with AI tools also remains an important issue. Differences in personal experience, career paths and working environments mean that not all lawyers have the same level of confidence or proficiency when it comes to these technologies. While financial barriers have been partially reduced by the emergence of more affordable and open models, disparities increasingly stem from differences in training, change management and organisational support.
Looking ahead to 2026, further developments are expected. One likely evolution is the growing role of AI agents capable of performing sequences of tasks with a higher degree of autonomy, under human supervision. Another promising area lies in the development of collaborative AI-enabled platforms, allowing lawyers and clients to work together more seamlessly on shared documents, analyses or workflows.
In this evolving context, investment in AI training remains a low-risk and high-value strategy. Understanding which tools to use, for which purposes, and with what limitations has become an essential component of professional competence. Legaltech, and generative AI in particular, should not be viewed as a threat, but as an opportunity to strengthen the legal profession in the service of clients, combining human expertise, critical judgment and technological efficiency.