The Legal and Compliance sector is witnessing a significant shift as AI-powered technologies promise to streamline operations, enhance decision-making, and reduce costs.
According to a recent Wolters Kluwer report, nearly three in four lawyers (73%) plan to integrate generative AI into their work within the next year, with nearly two-thirds (62%) already making changes to their day-to-day operations.
Despite the optimistic outlook, Bain’s Four Themes Emerging AI survey reveals that Legal professionals are the most disappointed by Generative AI, and that disappointment is increasing.
What might be driving this?
In short, most generic generative AI tools are not tailored for Legal and Compliance use cases, and their propensity for ‘hallucinations’ makes them unreliable for high-stakes work.
Over-advertised capabilities from new Legal AI vendors potentially fuel this disappointment, too, as exaggerated promises often fall short of actual performance. The recent, and somewhat controversial, Stanford University review of Legal AI tools is a good example of this.
Meanwhile, a UK Law Gazette report also cites concerns from Legal professionals over hallucinations (57%), security (55%), and a lack of trust in current free-to-use technology (55%).
In short, understanding these considerations is crucial in evaluating and adopting effective AI solutions — balancing the undeniable promise of AI with the practical realities of its implementation.
So what should Legal and Compliance professionals do when evaluating Legal AI solutions?
To find out, we spoke to a leader who has been through this process, Stefanie Briefs, Senior Legal Counsel and Lead of Legal Tech HR at Bertelsmann — one of the world’s largest media conglomerates with subsidiaries including Penguin Random House and Arvato.
But first, let us take a minute to quickly sum up where we’re at with Legal AI, and it’s potential benefits for lawyers.
AI tools for Lawyers
Almost every day, AI takes a big leap forward. Whether it’s creating highly-convincing travel videos, or summarizing complex 100-page reports, each day brings new capabilities that are impacting almost every profession. The same is true for Legal AI.
Most of the Legal AI solutions on the market today offer incredible efficiency savings that have an immediate impact on the productivity and profitability of law firms. The same can be said for in-house legal teams, with the additive incentive of making the wider business operationally more efficient.
What the future looks like is less clear. The potential is undoubtedly huge, that much is for certain. But as a cautious, compliance-led profession, we also need to be careful when evaluating and implementing Legal AI solutions to ensure they’re safe, secure and aren’t prone to hallucinations.
Benefits of using Legal AI tools in your law firm or in-house legal team
- Enhance Productivity: Automate routine and repetitive tasks with AI tools to streamline your legal team’s workflows.
- Save Time: AI processes and analyzes vast amounts of information in seconds, freeing up valuable time that you can devote to your clients or business.
- Improve Consistency: Minimize human error in repetitive tasks. AI ensures consistency in tasks like reviewing documents for inconsistencies or missing clauses.
- Elevate Client Experiences: With AI handling low-level tasks, law firms can focus more on complex legal work, enhancing client satisfaction, serving more clients, and fostering a client-centred practice. For in-house teams, you can also provide better services for your business and speed up wider business operations.
Key considerations when evaluating Legal AI solutions
1. AI has to be trained for Legal and Compliance use cases
To effectively serve the Legal industry, AI needs to be specifically trained on Legal and Compliance-related content.
General-purpose AI tools ultimately lack the nuance and specificity needed for legal applications, potentially leading to critical errors.
Training AI on legal content — contracts, laws, regulations, etc — will ensure it understands the context, terminology, and regulatory requirements unique to the field.
Similarly, training AI on Compliance policies and documents — internal policies, employee handbooks, etc — will ensure the AI understands the Compliance context, and can effectively support compliance professionals in their work.
Legal AI solutions need continuous learning from relevant legal texts, cases, and regulatory updates to stay current. They should learn from your own documents, too.
This ongoing adaptation helps maintain the relevance and accuracy of the AI outputs, making it a reliable tool for Legal and Compliance practitioners.
2. Answer quality has to be high, without errors or hallucinations
“The quality of the answers [that the AI delivers] has to be exceptionally high, with no room for errors or hallucinations,” says Stefanie Briefs, Senior Legal Counsel and Lead of Legal Tech HR at Bertelsmann
Legal decisions and considerations have to be based on accurate and reliable information. Any mistakes from an AI can have significant legal and financial repercussions, making it imperative that AI solutions deliver precise and trustworthy outputs.
The potential for AI to ‘hallucinate’ or generate plausible but incorrect information also poses a unique challenge in Legal — and should be considered accordingly. The Legal and Compliance AI solutions you evaluate need to be completely safe from hallucinations. Don’t just take vendors’ word for it either — test the solutions for yourself.
3. AI solutions have to be user-friendly to drive adoption
For an AI solution to be widely adopted within a Legal and Compliance setting, it needs to be user-friendly.
We’re not all tech experts, after all.
Any AI you evaluate needs to be accessible where you, your Legal and Compliance team work, and your business and employees work — whether that’s within Microsoft Teams, Slack, or your Outlook or Gmail inbox.
“A user-friendly solution with easily understandable responses is crucial for encouraging adoption and ensuring seamless integration into daily workflows,” Briefs notes.
After all, professionals are often pressed for time and may not have the technical expertise to navigate complex AI tools. AI interfaces should be intuitive, with minimal learning curves. Integration with existing software used by Legal and Compliance teams, and their business partners, will further enhance usability.
You should also look for a solution that understands and responds to queries in plain language so that it’s easy to interact with, without the AI needing to learn new commands or technical jargon.
Legal and Compliance professionals shouldn’t need to be “prompt engineers” to use an AI solution effectively.
4. Answers must have references to source documents
“All AI-generated answers need to include links back to the original source documents to ensure transparency and allow users to verify the underlying information,” emphasizes Briefs.
In regulatory practices, the ability to trace information back to its source is essential for verifying accuracy, establishing credibility, and building trust, not to mention additional research or context may be required from the underlying source document.
People using an AI solution to solve Legal and Compliance challenges, need to verify the sources of any AI-generated advice or documentation to ensure it is based on reliable and authoritative texts. By providing direct links to the source documents, advanced AI solutions can enhance transparency and accountability.
By allowing both Legal and Compliance professionals, and business partners, to instantly see and access the source texts, they can make more informed decisions, reduce risk, and have full confidence in using the AI.
5. The AI has to be secure and compliant with data privacy laws
Given the sensitive nature of regulatory data, any AI solution you evaluate needs to adhere to strict data privacy and security standards.
“The AI solution [you choose] has to comply with your stringent data privacy and security standards,” says Briefs.
Legal AI solutions must be designed to handle sensitive data in accordance with all the relevant laws and regulations, ensuring that information is protected at all times.
When you’re evaluating Legal AI solutions, ensure they’re SOC 2 Type II and ISO27001 verified, locally hosted, and GDPR compliant.
Final thoughts: AI needs to be built for Legal and Compliance
As AI continues to evolve, its integration into the Legal and Compliance sectors offers opportunities for efficiency and innovation. However, the transition must be approached with careful consideration of the unique demands and responsibilities inherent to Legal practice.
Stefanie Briefs’ insights underscore the importance of tailored AI solutions that are specifically trained for regulatory use cases, ensuring high-quality, accurate, and reliable outputs. User-friendliness and seamless integration into existing workflows are also crucial for driving adoption.
Moreover, the ability to reference source documents enhances the transparency and trustworthiness of AI-generated recommendations, while strict adherence to data privacy and security standards is non-negotiable given the sensitive nature of the data.
Ultimately, the successful adoption of AI in Legal and Compliance hinges on a balanced approach that aligns advanced AI capabilities with the practical realities and stringent requirements of regulatory practice.
By prioritizing these considerations, Legal and Compliance professionals can harness the power of AI to be more productive and take advantage of the latest advancements in technology to help achieve wider business objectives: whether that’s reducing risk, cutting costs or creating a better customer or employee experience.
