Four blind spots in AI legal advice

Some of the key risks of using AI in law are as follows:

1. Hallucination – this refers to the tendency of AI programmes to generate information that is incorrect, misleading, or entirely fabricated, while presenting it confidently and authoritatively.  Unlike a human lawyer, AI does not verify facts or know the law, which means it can invent case citations, attribute legal principles to the wrong authorities and mistake the outcome of cases.  A well-known case it which AI invented a case citation was Ayinde v London Borough of Haringey (2025), where the High Court identified that multiple authorities relied upon in judicial review proceedings simply did not exist which raised concerns that generative AI had been used in drafting.  The court was highly critical of those involved and referred both counsel and solicitors to their respective regulators for potential disciplinary action.
2. Outdated legal guidance can result in incorrect advice or flawed arguments as the law has evolved.  This is particularly problematic with AI tools, which may draw on older or incomplete data and fail to reflect recent developments in case law or changes in judicial approach.  Although not always immediately obvious, this type of error can weaken a party’s position just as significantly as fabricated authority, particularly where it is identified that the legal basis relied upon is no longer current.
3. Jurisdictional error – this is when AI produces legal analysis based on the wrong legal system.  Often, AI programmes can blend UK, US or other jurisdictions, resulting in incorrect procedures or legal principles being applied.   In the case of Ayinde (2025), as mentioned above, the court noted features in the submissions suggesting reliance on material not properly grounded in English law, as well as broader issues with inaccurate authorities.
4. Lack of context and nuance – AI programmes tend to generate generic, one-size-fits-all outputs which may overlook key nuances such as commercial objectives, risk appetite and the specific circumstances of the parties.  This can result in agreements that are technically correct but commercially inappropriate or unbalanced.  A human lawyer is therefore far better placed to ensure that the drafting reflects the parties’ specific facts and the realities of the market.

Whilst AI can undoubtedly assist with efficiency in legal work, it should not be relied upon as a substitute for professional advice.  As the risks above demonstrate, AI can overlook important details, misapply the law, or produce outputs that appear credible but ultimately create issues further down the line. A qualified lawyer is far better placed to ensure accuracy and tailor advice to the specific circumstances of each matter.  For reliable, commercially focused legal support, it is always advisable to seek expert guidance, and Greenwoods are well placed to support you in navigating these risks and getting it right from the outset.