2022 was a big year for digital contracting, with Ironclad AI transforming the way legal teams work every day. Advances in Natural Language Processing (NLP) allowed teams to speed up the entire contracting process, improve compliance, and flag problematic contract language.
So what about 2023? Are we full-on replacing lawyers with ChatGPT now? Not quite. But get ready for a year of excitement as the world of digital contracting changes once again: Get ready for the year of Generative AI and Large Language Models.
Large Language Models (LLMs) aren’t new to AI developers and researchers, but they’re newly poised to start shaping the way we work. An LLM is a type of machine learning model that can handle a wide range of NLP use cases. The most well-known LLM right now is OpenAI’s GPT-3 (Generative Pretrained Transformer 3). First released in June of 2020, GPT-3 is one of the largest and most powerful language processing AI models to date. The largest version of the model has roughly 175 billion parameters trained on a whopping 45 TB of text data — that’s roughly a half trillion words. It’s no surprise, then, that GPT-3 is widely considered the best AI model for generating text that reads like a human wrote it.
Two years after GPT-3 hit the Internet as a beta, a chatbot built atop the model went viral; ChatGPT took the mainstream Internet by storm in November 2022. The app racked up one million users in less than five days, showing the appeal of an AI chatbot developed specifically to converse with human beings. Though just a beta prototype, ChatGPT brought the power and potential of LLMs to the fore, sparking conversations and predictions about the future of everything from AI to the nature of work and society itself.
It’s no surprise, then, that the impact of LLMs dominates my list of AI predictions for next year. Simply put, 2023 will be the year of Large Language Models. Here are some thoughts on what’s ahead for AI in 2023.
2023 will be the year for Large Language Models
The ChatGPT craze has people dreaming about what might be possible with LLMs. Beta ChatGPT users have been asking the model to generate everything from school essays and blog posts to song lyrics and source code. Entrepreneurs, in turn, have been hastily cobbling together basic apps to start exploring ChatGPT’s power for a wide array of tasks. In early 2023, we’ll start to see those applications come to life, and start to change the nature of work.
Yet LLMs will be a disappointment for some in 2023
Important new technologies are usually ushered in with a bunch of not-so-important tries at making a buck off the hype. No doubt, some people will market half-baked ChatGPT-powered products as panaceas. LLMs are trained (in part) to give convincing answers, but these answers can also be untrue and unsubstantiated. Inevitably, some people will try to rely on these answers, with potentially disastrous consequences. Meaningful applications and advances built on the back of GPT-3 and other LLMs are just around the corner, to be sure. But beware the wave of less-than-great apps and business schemes bound to hit the market next year, alongside the good stuff.
But when applied correctly, LLMs will change the nature of work
This disappointment may temper some of the excitement and hype around LLMs, but the fact that LLMs are changing the nature of work will still be clear. Writing assistants, like Jasper and Copy.ai, will be the most obvious example, helping individuals and smaller teams quickly iterate on ideas and produce more content with fewer resources. Other document editors will follow suit), moving generative AI for language from the “early adopter” crowd to the “early majority” crowd.
Non-generative AI will quietly advance
Less obvious will be the influence on non-generative applications of LLMs. Text classification and named entity recognition (NER) — currently used for tasks like extracting information from large amounts of text — will noticeably improve, enabling a much wider array of applications.
Mass job elimination? No, but…
LLM-powered bots aren’t going to displace thousands of writers and content developers en masse next year. But foundation models will enable new challengers to established business models. In media, small outfits will be able to produce high-quality content at a fraction of the cost (consider Stable Diffusion for image generation, for example). Similarly, small, tech-enabled legal practices will start to challenge established partnerships, using AI to boost efficiency and productivity without adding staff. Artificial intelligence will act as our co-pilot, making us better at the work we do and freeing up more time to put our human intelligence to work.
Many of these amazing advances in AI will become pedestrian
When voice recognition like Siri came out in 2011, we couldn’t believe it – it was a miracle of technology that felt like magic. But today, people hardly even bat an eye as they pull out their smartphone to ask where the closest pet store is. In 2023 and beyond, we’ll begin to see commercial and business uses of AI that will undoubtedly be woven into our lives on a daily basis. We will look at something like generative AI and not even remember the “before times.”
2023 Will Be a Year of Milestones for Contract AI
When thinking about what’s ahead for AI in 2023, Generative AI will no doubt change the way we do lots of things, including work. Don’t expect ChatGPT to be representing you in court next year. But by the end of 2023, LLMs will certainly have changed the way people create, share, and use digital contracts.
In fact, we’ve got some *super* exciting enhancements for Ironclad AI coming up soon. Follow us on LinkedIn, Twitter, and Instagram to make sure you’re the first to know – or sign up for a custom demo now to see how Ironclad AI can help you speed your business, drive results, and operate more effectively.
Ironclad is not a law firm, and this post does not constitute or contain legal advice. To evaluate the accuracy, sufficiency, or reliability of the ideas and guidance reflected here, or the applicability of these materials to your business, you should consult with a licensed attorney. Use of and access to any of the resources contained within Ironclad’s site do not create an attorney-client relationship between the user and Ironclad.
