Europe’s startups have been riding a wave of innovation, creating numerous applications of generative and agentic AI. This innovation has explored the art of the possible, but it has also flooded the world with slop. Content is awash with it, whether it is LinkedIn posts or quotidian corporate drivel. The slop extends to legal work AI was supposed to make work easier, but now many are grappling with what Jamie Bartlett, author and BBC journalist, describes as “the burden of AI clog“. Cleaning up this mess means that AI is not saving time, but is actually creating more work. The slop extends to the legal domain, with contract negotiations and other legal work now being absorbed by a lengthy review process because of AI. Even the bodies that are supposed to be drafting the laws are swimming in slop. In the US, for example, the Office of the Legislative Counsel, is reportedly overwhelmed by the error-riddled drafts that AI has generated. The problem isn’t AI per se Advances in LegalTech in Europe, and beyond, means there are many tools to assist with legal work. Despite this, however, their output is of varying quality. And some law firms still don’t have a proper policy in place for how employees use AI, which means some teams are relying on general LLMs (like ChatGPT or Claude) for their drafting and reviewing. As contracts are amended in negotiations, much of the document is struck through and sent back saturated with red lines. AI editors are ripping and replacing much of the content, and the changes usually lack insight and relevance. This is hardly the way to engage in a negotiation, and the trigger-happy nature of the tools is likely to aggravate the other party. Judgement and discretion have been lost in the process. Many tools don’t know the difference between dogmatically applying the letter of the law and letting things go for the sake of progressing a negotiation. Unnecessary redlining can be stopped This, however, is a problem that Europe’s startups can solve. There are ways to ensure that legal work doesn’t descend into slop, and to prevent the revisions turning into an infinite ping pong match between AI agents. The focus instead needs to be on judicious editing, where the tool knows when to leave something alone. The amendments should be minimal, and these should be aligned to a well-defined playbook. This playbook sets out a firm’s position, policies and approach and sets parameters for what is, and isn’t, acceptable. The tool scans for compliance issues, conflicts and highlights areas of difference. It can then make suggested edits that align with this well-defined playbook and offer a best-case scenario for the company, or firm. This approach is more likely to bridge the gap between the parties and get to a consensus quicker, without the need for human intervention. AI should be a blessing, not a burden When used correctly, legal AI can be a blessing for companies and means more work can be done in-house before calling in outside counsel. Such tools can be used as mediation portals, where differences of position are highlighted and evaluated first instead of text being ruthlessly red-lined. The excessive redlining needs to stop, as does the AI slop and the AI clog burden it creates. At the moment, users are putting up with it albeit complaining about the extra work that AI creates. Soon, however, the slop won’t be tolerated. We expect the bar to be raised of what AI-enabled work should look like. Expectations will shift and users will soon demand AI tools that edit in a way that is precise, relevant and insightful. The post OPINON: Europe’s legal work is awash with AI slop, but there is a way to fix it appeared first on EU-Startups.
