2025 is a bit of a fortune teller’s playground. You almost can’t go wrong by making a prediction because there’s so much fertile ground for growth and change in the tech sector this year. Will this be the year of robotics? Will there be a chip shortage? Will AI finally have feelings? In this article, we’re going to look at what we’re anticipating for the legal tech sector in 2025 and where we’re hoping things will change.
1. Discoverability of Transcriptions. If you haven’t experienced AI’s power to transcribe, summarize, and draft documents yet, you soon will. We’ve seen a few legal AI apps in action, and we predict there will be a multitude sooner rather than later. Overall, the transcribe and summarize functions are already going mainstream in AI notetaking applications. Apple Intelligence, now available on various device models with iOS 18.1 and later, will perform the transcription and summarization service on calls recorded via the Phone app and audio recordings created with the Notes app. With the variety of laws already on the books regarding the legality and notification requirements for phone call recording, we also predict that there will be some future legal wrangling about whether the transcriptions and summaries of these calls on those devices should be part of discovery.
2. Drafting Pleadings with AI. AI’s ability to shorten the time required to draft pleadings is so well known that unless we were making a contrarian prediction, we wouldn’t bother predicting any fall off for this trend. What we hope to see decrease are hallucinations and missummarizations in draft pleadings. While the news about hallucinated citations is well-known, what you may have missed is a warning that AI summaries of court orders may erroneously provide an incorrect conclusion because the tool may not properly digest the entire order. Regarding summarization, we predict that AI developers will be working to increase reliability in 2025.
3. AI Tools for Document Review. In line with our first two predictions above, AI used for document review will increase. The short amount of time required for an AI tool to review massive amounts of data is too attractive for firms and developers not to continue pursuing. However, what we predict is that pricing models encompassing credits and tokens using “pay to play” tools will proliferate.
4. Hyperlinked Files. We’ve been tracking the development of hyperlinked files or modern attachments as an eDiscovery issue since the Lotus Notes days, so we believe that we’re going to continue to see more ESI protocols regarding hyperlinked files throughout 2025. Along with that, we expect to continue seeing the data explosion that comes hand-in-hand with collecting hyperlinked files which can be voluminous. A number of prior touchstone cases (In re: Uber, Nichols v. Noom, In re StubHub and more) raised questions regarding hyperlinked files, and our reading of the issue is that it’s not settled yet. 2025 will continue to see courts weighing the pros and cons of collecting, processing, reviewing, and producing massive data sets with hyperlinked files.
5. Deepfakes. If you think “we’re there yet” with deepfake technology, think again. Deepfakes, be they video, audio, or static images, are still upleveling in quality and creation speed. Gen AI leaps in image and audio creation reverberate rapidly to the deepfake realm. And where there are deepfakes, there are scams. So, while we predict that deepfake scams will increase in sophistication, variety, and reach, we’re also predicting that 2025 will be the year that watermarking AI-generated content becomes more available and popular in the fight to authenticate content, protect IP, and fight scammers. We may even see more legislation designed to deter deepfake misuse.
The fun of predictions isn’t in making them so much as seeing how the future unfolds. And while that’s amusing, we also take our predictions seriously so that we can better anticipate what our clients will need in the eDiscovery, digital forensics, and cybersecurity markets over the coming year.
