A Lawyer's Perspective on AI and the Future of China's Legal Services
1. Introduction: Thoughts on legal practice under the wave of technology
As a lawyer who has worked for many years, I deeply feel that artificial intelligence (AI) is having an unprecedented impact on our long-standing and rigorous industry. This is not a distant idea, but a technological reality that has already penetrated our work environment and begun to gradually change the practice model. From the initial discussion and careful observation of automated legal services to the current emergence of various "smart judicial" applications, we are in a period where technology is profoundly driving changes in the industry.
This round of change benefits from the rapid iteration of artificial intelligence technology on the one hand, and it is also inseparable from the strategic deployment and policy guidance at the national level. The construction of "smart rule of law" has been included in the top-level design, which has undoubtedly accelerated the application and exploration of AI technology in the field of legal services. From judicial agencies to law firms and corporate legal departments, the introduction of AI technology is discussed or practiced to varying degrees in all aspects.
However, amid the upsurge of technological development and policy promotion, as front-line legal practitioners, we are more concerned about: What is the specific impact of AI technology on daily practice? Is it an effective tool to improve service quality and efficiency, or does it bring new practice burdens and potential risks?
This article aims to analyze the actual application status and challenges of AI technology in my country's legal services field based on the perspective of a practicing lawyer, combined with personal observation and industry practice, and explores how legal professionals should seek new paths for professional development in this context of change.
2. Rules first: prudently explore AI applications within the legal and policy framework
Under China's special legal and policy environment, the application of any new technology must be premised on understanding and following relevant rules, especially the application of artificial intelligence. The country’s strategic plan for AI development and the grand blueprint for the construction of “smart rule of law” have set the tone for the industry: innovative development is encouraged, but it must operate within a standardized and controllable track. For practicing lawyers, this means recognizing the opportunities presented by technology while also maintaining a heightened awareness of compliance.
(1) A dual-track strategy of parallel development and supervision
At the national level, through programmatic documents such as the "New Generation Artificial Intelligence Development Plan", the ambitious goal of promoting technological innovation has been clarified. At the same time, regulatory agencies also quickly issued a series of targeted normative documents aimed at guiding and restricting technology development in the early stages. This forward-looking regulatory strategy is leading in the world.
For example, the "Regulations on the Management of Algorithm Recommendations for Internet Information Services" stipulates that the use of algorithms to influence information distribution (such as case push, risk assessment tools) must follow the principles of fairness, transparency, and value orientation. This has triggered our reflection on the objectivity of the legal technology products we rely on. The "Internet Information Services Deep Synthesis Management Regulations" directly respond to the possible impact of deep forgery technology on evidence rules, reminding us that we may need to have the ability to identify the authenticity of AI-generated content in future practices, and be wary of the improper use of related technologies in our work. The "Interim Measures for the Management of Generative Artificial Intelligence Services" are more closely related to our daily work. They impose multiple obligations on AI service providers such as the legality of data sources, content accuracy, and security assessments. This means that when we choose AI tools, in addition to functionality, compliance review has become an indispensable consideration.
It should be recognized that although this model of “developing while regulating” reflects the proactive nature of supervision, it also brings adaptability challenges to the market. The rapid updates of regulations require that the tools we use and the workflows we establish must remain flexible to meet new compliance requirements at any time.
(2) Data foundation and compliance boundaries of AI: Data compliance is the lifeline
The effectiveness of artificial intelligence is highly dependent on data, and data in the legal industry is highly sensitive. We, the lawyer community, are deeply aware of this.
First of all, the Cybersecurity Law, Data Security Law and Personal Information Protection Law together form the legal bottom line that must be observed when processing any data (including training AI models or using AI to process customer information). Client entrustment is based on trust, and protecting the confidentiality of client information is not only a core professional ethical requirement, but also a legal obligation.
Secondly, the "Interim Measures for the Management of Generative Artificial Intelligence Services" put forward clear requirements for the legal sources of training data, intellectual property protection, and personal information protection. This is directly related to the compliance risks of our choice of AI tools. A practical concern is: If an AI tool uses illegally obtained or unauthorized data for training, will we bear associated compliance risks when using this tool?
In addition, several issues in the field of intellectual property remain to be clarified. Does AI training data constitute an infringement of existing copyrights? There are no unified and clear rules around the world regarding the legal attributes, copyrightability and ownership of rights of AI-generated content (AIGC). Although China has judicial precedents that have initially recognized the copyrightability of AIGC under certain conditions, the relevant rules are still in the process of being formed. This means for our practice that when using AI to assist research or document drafting, we must be extremely cautious to avoid inadvertently infringing on others’ intellectual property rights or making improper judgments about the ownership of rights to AI-generated content.
For example, if a contract clause drafted using an AI tool is highly similar to a copyrighted template, does this constitute infringement? For another example, when using AI to analyze evidence materials submitted by the other party that may contain trade secrets, how to ensure that the entire process complies with confidentiality obligations and data security regulations? These are real issues that we must conduct refined risk assessment and trade-offs in our daily practice.
(3) Ethical considerations: value orientation and professional ethics constraints in technology application
In addition to laws and regulations, professional ethics are also important guidelines for guiding the application of AI. Relevant regulations have repeatedly emphasized that AI applications must abide by social morality and ethics, adhere to correct value orientation, prevent discrimination, and ensure the safety and controllability of technology applications. the Supreme People's Court also particularly emphasizes the principle of ethics first when promoting the judicial application of AI.
For lawyers, ethical norms are not only external constraints, but also internal requirements for our professional ethics. The introduction of AI technology must not shake our fundamental responsibilities such as safeguarding fairness and justice, protecting vulnerable groups, and adhering to professional confidentiality. For example, over-reliance on AI that may contain biases for case evaluation may affect the objectivity of the professional opinions we provide to clients, and may even exacerbate existing injustices in society. This is the red line in practice that we must always be vigilant about.
All in all, my country’s regulatory framework for AI applications in the legal field is gradually becoming clearer, showing the characteristics of “strong guidance, quick response, and emphasis on compliance.” This delineates clear action boundaries for practicing lawyers, and also requires us to maintain a high degree of industry sensitivity and continuous learning capabilities, so that we can prudently explore and utilize the potential value brought by AI technology while ensuring compliance.
3. The practical temperature of AI in the legal profession: from market popularity to differences in individual feelings
Although positive signals continue to be released at the policy level, market discussions on legal AI are also very active. Various research reports predict huge market potential, and new tools and platforms are constantly emerging. However, as a practicing lawyer, I feel that the "temperature" of AI applications is not generally high, but shows a significant "unevenness" phenomenon.
(1) Market Observation: Precipitation and pragmatic considerations in the hustle and bustle
It is undeniable that investment and entrepreneurial activity in the legal technology field was once active. We have witnessed the continued deepening of established manufacturers such as France-Italy Technology, and we have also seen emerging platforms such as iCourt Alpha and Panda AI attract market attention with their AI concepts. Even official research institutions (such as China Legal Research) and top universities have also made arrangements.
However, judging from the daily work perception of practicing lawyers:
1. The distance between “applause” and “popularity”: Despite the loud marketing, there still seem to be a handful of AI tools that have been integrated frequently and deeply into the workflow of lawyers. In most cases, we may only use it on a trial basis in specific links (such as preliminary information retrieval), or participate in relevant training lectures, and there is still a long way to go before it becomes an indispensable core productivity tool.
2. Prudent evaluation of input-output ratio: Law firms, especially small and medium-sized law firms, usually have a very pragmatic attitude when introducing AI tools. The high purchase or subscription cost, coupled with the difficulty in accurately quantifying its performance improvement (for example, it is often difficult to accurately calculate how much loss an AI contract review tool can recover or how many man-hours it saves) makes the investment decision-making process extremely cautious. In contrast, although the "soft" returns brought about by efficiency improvements are attractive, they are often less urgent than direct cost control pressure.
3. The problem of building a trust foundation: The core of legal services lies in the trust relationship. We generally have deep concerns about the accuracy and reliability of the output results of AI tools, especially the confidentiality when handling sensitive customer information. A serious error or information leakage can cause irreparable damage to a law firm's reputation and client relationships.
(2) AI application portraits of different legal entities: significant temperature differences
Further observation reveals that there are huge differences in the degree of adoption of AI technology among different types of legal service entities.
1. Large law firms: Head law firms usually have stronger financial strength, strategic vision and technology integration capabilities, so they are more inclined to invest resources in informatization and intelligent construction. They may choose to independently develop or customize AI tools for use in internal knowledge management, contract review, assisted due diligence and other scenarios. Even so, as far as I know, AI usually plays a supporting role in these institutions, rather than replacing lawyers in core judgments.
2. Small and medium-sized law firms and individual lawyers: These are the main groups of lawyers in China. AI is more of an “option” than a “standard configuration.” Cost-effectiveness, ease of use and practical results are the primary considerations. Many peers may still mainly rely on traditional legal databases, or try to use some free or low-cost general-purpose AI for auxiliary writing or information collection, but their output results will be subject to extremely strict verification and verification.
3. Corporate legal department: Corporate legal affairs, especially the legal teams of large enterprises, are often active explorers and promoters of the application of legal technology. They face dual pressures to control external legal service costs and improve internal work efficiency. Therefore, in areas with a relatively high degree of standardization such as contract management and compliance review, their acceptance and willingness to invest in AI tools may be relatively higher. Some companies have even begun to explore the deployment of localized AI solutions to strengthen data security.
4. Judicial agencies: undoubtedly the “high ground” for current AI technology applications. Driven by policies such as "Smart Courts" and "Smart Procuratorate", courts and procuratorates have systematically deployed a large number of AI applications, covering multiple judicial links from case filing, trial assistance, automatic document generation, similar case push to prosecutorial supervision. This kind of top-down large-scale application means that for us practicing lawyers, we need to adapt to new modes of interaction with these intelligent systems (such as electronic files, online court hearings, etc.), and it may also indirectly affect our selection and evaluation standards of commercial AI tools.
(3) Overview of AI Tools: Selection Considerations and Practical Challenges
Facing various legal AI tools emerging on the market, as legal professionals, we need to make prudent evaluations and choices, not only to understand their potential value, but also to clearly understand their limitations and risks. The following is an analysis of several representative tools:
1.iCourt Alpha System
Positioning: Committed to building a "one-stop" legal service platform with a wide range of functions and designed to integrate lawyer workflow.
Advantages: In theory, it can reduce transactional work time and realize process integration management.
Challenge: Functional comprehensiveness may come at the expense of depth and precision at a single point of functionality. There is still a gap between the quality of documents, the accuracy of analysis, and the reliability of risk assessment compared with experienced lawyers. The cost is high, and the law firm needs to carefully evaluate the input-output ratio.
2. Legal research open platform
Positioning: Relying on official background to provide authoritative judicial data resources and analysis functions.
Advantages: The data is highly authoritative and suitable for basic legal research, case retrieval, and trend analysis. After combining with large models, it has great potential in extracting complex case elements and mining judgment logic.
Challenge: The functions may focus on macro data analysis, and may be far from the detailed needs of lawyers for specific case handling. User interface and interaction logic may need to be adapted.
3. Fayi Technology
Positioning: Deeply engaged in the informatization of the judicial system, with deep technical accumulation in judgment document processing and judicial data mining.
Advantages: Solid technical foundation, profound understanding of judicial business, and strong processing capabilities. Possible access to deeper structured judicial data.
Challenge: Its core advantages may focus on serving the judicial authorities (B2G), and the maturity, ease of use and fit with lawyers' workflow for lawyer-oriented (B2B/B2C) products are key.
4. Panda AI
Positioning: Emphasize "credibility" and "data quality", and provide accurate and reliable legal information and analytical support.
Advantages: With "credibility" as the core selling point, it addresses the core concerns of lawyers and helps build user trust. Focusing on core functionality can lead to deep advantages.
Challenge: As a relatively new market player, its advantages need to be widely verified by the market. Brand awareness and user base still need to be accumulated.
5. General-purpose large language models (such as ChatGPT, Gemini, DeepSeek, Doubao, etc.):
Positioning: General-purpose AI, designed in non-legal vertical fields, with powerful natural language processing capabilities.
Advantages: Easy to obtain and low threshold for use. It can provide assistance in information retrieval, preliminary concept explanation, drafting of non-core documents, brainstorming, etc., with high efficiency.
Core risks:
(1) Lack of legal professionalism: The lack of accuracy in citing legal terms, principles, and cases is its "flaw".
(2) "Illusion" problem: It is extremely easy to generate false information (such as fabricating laws and cases), which is never acceptable in serious legal work.
(3) Timeliness and accuracy are questionable: the update of the knowledge base may lag behind, and the information may be outdated or wrong.
(4) Confidentiality risk: Entering confidential information involves significant risks of data leakage and violation of professional confidentiality obligations.
Choosing an AI legal tool is not a simple matter of technology adoption, but requires a comprehensive assessment based on business needs, team size, cost budget and risk tolerance. There is no "panacea", only "applicable products". More importantly, no matter what tools are used, the core values of lawyers - independent professional judgment, rigorous logical reasoning, loyalty obligations to clients and adherence to the spirit of the law - are always irreplaceable. AI should be used as an auxiliary means to improve our professional capabilities, rather than replace our independent thinking.
Currently, the application of AI in the legal industry is still at the stage of “huge potential, but prudent implementation”. A large number of basic legal service needs have yet to be met, and within the lawyer community, how to effectively control AI, maximize its value, and avoid potential risks is still an important issue commonly faced by lawyers.
4. Integration of AI into practice: from efficiency improvement to exploration of human-machine collaboration model
In the daily practice of lawyers, AI is being introduced at different levels. According to my observation, its application can be roughly divided into several levels, but no matter which level, the current mainstream is not intended to replace, but to explore how humans and machines can work together more effectively.
(1) Level 1: AI as an “advanced information processing assistant” – automating basic work
This is the most widely used and least controversial aspect of AI. Its core goal is to liberate lawyers from matters that are time-consuming, highly repetitive but require relatively low professional judgment.
Typical application scenarios include:
1. Intelligent search optimization: Compared with traditional databases, it may provide faster and more comprehensive search results for regulations and cases, significantly saving screening time.
2. Document management assistance: Automatically classify and label large amounts of evidence materials or project documents to improve sorting efficiency.
3. Transcription of court hearing records: Speech recognition technology can theoretically assist in court hearing recording, but in practice, due to factors such as dialect, speech speed, noise, etc., the accuracy still needs to be manually verified.
4. Basic template calling: matching contract or document templates based on simple elements, but there is still a long way to go before the true "legal document drafting".
The value of AI at this level is that it can indeed save time and reduce the probability of some low-level errors. But its limitations are also very obvious: it is more like a "toolbox" with upgraded functions and does not touch the core judgment link of legal work. The time thus saved often needs to be invested in verification, verification and deeper analysis and judgment of AI output results.
(2) Level 2: AI serves as a “junior professional assistant” – assisting core legal business
At this level, AI has begun to try to intervene in some substantive legal work, but it is still in a supporting position, and its output must be strictly reviewed and controlled by experienced lawyers.
Typical application scenarios include:
1. Intelligent contract review: AI can quickly scan the contract text to identify standard clauses, missing elements, or potential risk points. This has some value for preliminary screening or processing of large numbers of standardized contracts. However, it is difficult for them to deeply understand the complex transaction background, identify the hidden business risks, or grasp the delicate balance in negotiations.
2. Assisted generation of documents: Based on structured element input, AI can generate the first draft of litigation documents, lawyer's letters, etc. This can be used as a starting point for the work, but it still requires a lot of revision, polishing and reshaping of legal logic by lawyers before it can form a qualified legal document that is logically rigorous, persuasive and meets the needs of the individual case.
3. Preliminary screening and sorting of evidence: In scenarios such as electronic evidence discovery, AI can assist in screening and classifying massive documents. But this still requires lawyers to preset precise screening criteria and manually review, interpret and judge the key evidence that has been screened out.
As a "junior professional assistant", AI can improve the work efficiency and standardization of specific links. But the core risk lies in over-reliance. AI may not fully understand the nuances of a specific context, its algorithms may have built-in biases, and it may even produce misleading information. If a lawyer adopts his suggestions or generated text without critical review, it may lead to serious consequences and even constitute professional negligence. This is a risk point that must be highly vigilant in practice.
(3) Level 3: AI as an “analysis and prediction tool”? : : The temptation and practical challenges of advanced applications
This is the "frontier exploration area" of AI applications, trying to simulate lawyers to conduct more complex analysis, predictions and even provide strategic suggestions. Currently, it is mainly in the exploratory stage, with extremely limited practical applications and high risks.
Main exploration directions include:
1. Case outcome prediction: By analyzing a large amount of historical judgment data, trying to predict the winning probability or sentencing range of specific types of cases.
2. Judicial tendency analysis: Analyze the judicial style and preferences of a specific judge or court to provide reference for formulating litigation strategies. Such applications are extremely attractive, but their reliability and ethical boundaries are problematic.
3. Simulated legal reasoning: The academic community is trying to use AI to simulate court debates and explore the possibility of complex legal reasoning.
Although these directions sound promising for “intelligence”, real-life applications face severe challenges:
1. Data quality and completeness are fundamental constraints: the accuracy of prediction is highly dependent on the quality and representativeness of training data. Public judgment documents are often only the "tip of the iceberg" of trial results and cannot fully reflect all considerations in the trial process. Using biased or incomplete data to make predictions raises questions about the reliability of the results.
2. The "black box" problem of algorithms is difficult to solve: the specific reasoning process by which AI reaches prediction conclusions is often opaque. This is in fundamental conflict with the clear logic and rigorous argumentation required by the legal profession. Lawyers cannot effectively explain an “AI prediction” based on an unexplainable algorithm to clients or the court.
3. The inherent complexity of law is difficult to fully model: legal practice is not a simple mathematical operation. It is deeply embedded in complex factors such as value judgment, the integration of emotion, reason and law, social experience and even the individual cognition of judicial personnel. These "humanized" and unstructured elements are difficult to accurately capture and simulate with current AI technology.
4. The risk of “illusion” is a fatal flaw: If false information generated by AI (such as incorrect references to legal provisions and fictitious precedents) is misused for case analysis or strategy formulation, the consequences will be disastrous.
Therefore, for practicing lawyers, the current use of AI for advanced analysis and prediction should be seen as providing a "reference perspective" or "data insight" rather than a basis for decision-making that can be directly relied on. Any such output can never replace a lawyer's independent professional judgment based on solid professional knowledge, rich practice experience and an in-depth understanding of the specific circumstances of the case.
(4) “AI + Lawyer”: the most realistic human-machine collaboration model currently
Taken together, at present and in the foreseeable future, the role of AI in China's legal practice is clear: it is a tool and assistant to assist lawyers in their work, not a substitute. We are gradually forming and deepening an "AI+lawyer" human-machine collaborative working model:
1. Clear division of labor: AI is responsible for processing massive amounts of information, identifying data patterns, generating preliminary texts, and improving basic work efficiency; lawyers are responsible for defining core issues, grasping the direction of the case, reviewing and verifying information, making complex legal judgments, formulating litigation or non-litigation strategies, communicating and coordinating with all parties, and ultimately assuming professional responsibilities.
2. Adaptation and reshaping of lawyer skills: Lawyers need to learn how to “interact” with AI effectively. This means mastering:
(1) Precise instruction ability: How to issue clear, specific, and effective instructions to AI tools to guide them to produce high-quality preliminary output that meets needs.
(2) Critical evaluation ability: Maintain a high degree of professional vigilance for any output of AI (whether it is search results, contract clause suggestions or preliminary drafts of documents), and use your own legal knowledge and experience to review and verify its accuracy, compliance, applicability and potential risks. This is the core competency of lawyers and the last line of defense to ensure service quality.
(3) Integrated application capabilities: Integrate AI tools organically into existing workflows so that they can truly serve the goal of improving professional services, rather than being formally “used for the sake of use.”
As AI gradually takes over more basic and repetitive tasks (especially tasks that may have been performed by junior lawyers or paralegals in the past), the value chain of the legal services industry is undergoing reshaping. In the future, lawyers' core competitiveness will be more reflected in areas that are difficult for AI to reach: in-depth legal logical thinking ability, the ability to creatively solve complex legal problems, accurate risk prediction and management capabilities, high-level communication, negotiation and persuasion skills, keen business insights, firm professional ethics, and the wisdom to harness emerging technologies to empower professional services. The transformation from the traditional "legal craftsman" to the role of "legal advisor + strategist" has become an inevitable trend in the development of the industry.
5. Future Direction: Embrace change, adhere to values and move forward pragmatically
Facing the profound changes brought about by artificial intelligence, we lawyers can neither rest on our laurels nor be overly anxious. The key is to clearly understand development trends, adjust practice strategies when appropriate, proactively adapt to changes, and adhere to our core values as legal professionals in the wave of change.
(1) Foreseeing the future: several trends in the development of legal AI in China
Combining current industry dynamics and technological potential, I make the following preliminary judgments on the development trends of China’s legal AI field in the next few years:
1. The market will continue to develop, but will become more rational and pragmatic: Capital and users will pay more attention to AI tools and solutions that can effectively solve industry pain points, bring clear business value, and have relatively reasonable input and output.
2. The human-machine collaboration model of "AI + lawyers" will dominate and deepen in the long term: AI's role as lawyers' "intelligent assistant" and "analytic amplifier" will be further strengthened, rather than replacing the core functions of lawyers on a large scale.
3. Large legal models in vertical fields are the key breakthrough direction: vertical large models with in-depth training on China’s legal system, judicial practice and specific business fields will become the core competitiveness of legal technology companies.
4. Explore more intelligent application scenarios, but remain cautious: You may carefully explore the application of AI in simulating legal games, assisting complex decision-making, and conducting deeper legal reasoning on the premise that risks are controllable.
5. New legal service models driven by technology may emerge: Especially in the field of legal services with a high degree of standardization, alternative legal service providers (ALSPs) based on AI technology may gain more room for development.
6. Compliance, security and trustworthiness have become core competitiveness: With the improvement of supervision and the improvement of user awareness, the ability to provide compliant, safe and reliable AI products and services will become the "lifeline" for the survival and development of legal technology companies.
(2) The evolution of lawyers: Reshaping and strengthening core competitiveness in the AI era
Artificial intelligence will not make lawyers unemployed, but it will undoubtedly accelerate the elimination of lawyers who cannot adapt to the changes of the times and fail to enhance their own value. Therefore, we need to proactively evolve, reshape and strengthen our core competitiveness:
1. Improve technical literacy instead of pursuing becoming a technical expert: You need to understand the basic principles, capability boundaries, and potential risks of AI, and master the skills to effectively use relevant tools, especially the ability to ask precise questions and critically evaluate.
2. Return to and strengthen core legal skills: The more AI can handle standardized tasks, the more important it becomes for lawyers to have in-depth legal thinking, complex problem analysis, creative solution design, and strategic legal planning.
3. Deeply cultivate professional fields and build professional barriers: Developing profound professional knowledge and practical experience in specific legal fields and providing highly specialized and refined services are effective ways to cope with homogeneous competition and technological impact.
4. Amplify "humanized" value: Improve empathy, communication skills, trust building, ethical judgment, emotional value provision and other "soft skills" that are difficult to replicate for AI. These are the keys to maintaining customer relationships and handling complex interpersonal interactions.
5. Embrace lifelong learning and maintain professional acumen: The law itself is constantly evolving, and technology is also rapidly iterating. Maintaining an open mind and continuing to learn new legal knowledge, industry trends and technical tools are the basis for lawyers to remain competitive.
6. Conclusion
The integration of artificial intelligence technology and Chinese legal practice is an ongoing profound and complex change. The macro guidance of the national strategy, the pioneering practice of the judicial system, and the gradual emergence of market demand jointly paint a picture full of opportunities and challenges.
As practicing lawyers caught up in this torrent of change, we not only see the potential of AI in improving work efficiency and assisting professional judgment, but we are also deeply aware of its inherent limitations in terms of reliability, security, and replacement of core legal thinking. The "AI+lawyer" human-machine collaboration model will undoubtedly become the mainstream form in the field of legal services for a long time to come.
Ultimately, the value of technology lies in how it serves people and the ultimate goal of the rule of law. In this industry revolution driven by AI, we lawyers must always adhere to the bottom line of professional ethics and the core of professionalism, ensuring that technology is always used as a tool to achieve justice and improve services, rather than the purpose itself that transcends legal principles and human judgment.
When AI can generate hundreds of contracts in one minute, the core value of lawyers is no longer "speed" but "accuracy"; when AI can list thousands of cases, lawyers' winning hand is no longer "memory" but "insight". Technology will eventually reshape the form of legal services, but the warmth of the rule of law will always lie in people's belief in justice.