Are fears around AI getting in the way of innovation and evolution?

Looking over the recent Definely survey on LegalTech adoption, one stat really stood out to me. While LegalTech and Automation are the main areas of investment for firms and legal departments (making up almost half of all LegalTech investment between them), 33% of firms have no plans at all to invest in AI or automation. 

Not all AI investment is a guaranteed winner, obviously - 24% have said they're not satisfied with their LegalTech investments. But over a quarter of respondents said that they are 'significantly better off' for using AI (compared to just 2% saying they're significantly worse off) - so those that are finding the right tools are seeing significant advantages.

In LegalTech, those advantages include more efficiency, better accuracy and more productivity. Or, to put it another way, creating space for thought, not toil. So why are almost a quarter not satisfied?

Considering those that are getting it right are seeing significant improvements, and the efficiency on offer, the main question that jumps to mind is ‘why would you not invest?’

Fears over LegalTech AI and security

One of the biggest areas of concern - 58% overall - is security, around training data, transparency around how AI finds results, and the fears of security leaks.

The concern is understandable. Legal data, whether commercial, personal or intellectual property, is something no organisation wants to risk. Even aside from the massive area of reputational damage, the legal issues involved in data breaches are enough to cause serious problems for any firm. 

Put simply, nobody wants AI trained  on their data if there’s any chance that data might fall into the public domain or benefit competitors. This is the stuff of data nightmares.

But this isn’t an issue that’s going to be helped with a lack of awareness or understanding - in fact, 30% of respondents said their lack of understanding of AI was their biggest area of concern. And that’s important, because it can lead to a fear of what people think AI is or a misunderstanding based news stories around training data and intellectual property - whereas vendors that are building AI tools for specific use cases (like legal workflows) are a different game entirely.

Are you considering solutions that already exist?

While security and privacy are understandable concerns, what steps are organisations taking to actually solve it? Because solutions exist and are continually being improved, and they open up the advantages that one in four companies are finding from AI tools.

It’d be interesting to understand how many companies would consider hosting in their own private cloud to overcome this issue, for starters. Or how many would consider using third-party hosting with a known or trusted vendor.

With cloud space available from companies like AWS, Azure or Google, you’ll also benefit from their technological experience and support. The amount of staff they have working on data security alone means that their understanding of ongoing cybersecurity is unparalleled.

The odds are that private-cloud-based data storage will be more secure than it would be in a private server (especially if it’s a self-hosted centre that you don’t actually own anyway).

If you haven’t already considered this, it might be worth reflecting if this is an approach that could open up the world of AI and LegalTech to you. It’d also be dealing with the reality of security rather than the fear of it.

Security-conscious AI approaches

If you’re worried about your own lack of understanding, taking on some training on AI could help a lot, not least in helping you separate out the vendors who can actually help you from those promising more than they can deliver. 

As with many other areas, LegalTech has also attracted developers who try to adapt their offerings to specific problems, without understanding the complexity the industry involves - so it’s important to avoid working with vendors offering nothing more than a house of cards.

Finding the right vendor involves the same kind of rigorous checking as you would with employees. In the same way as you’d do background checks before giving someone access to sensitive information, you need to include vendor security protocols as a consideration before working with them. It’s KYC but for vendors. 

This greater understanding can also help with accuracy, the biggest area of concern of them all, according to our respondents. By working with appropriate tools and using them well, you’ll be likely to see better results than less legal-specific tools. I recently wrote about how, AI should augment lawyers, not replace them. It's important to understand where AI works well and where it doesn't. Otherwise, you can rely on it for the wrong things, and find it difficult to properly balance where human checks and changes are critical.

Collaboration and competition

On an industry level, firms are going to need a broader horizon view if LegalTech is going to properly evolve and take root in our industry. 

There can be an attitude around co-building about not wanting to build tools that will then benefit competitors. But if we want people to build more products in this industry to an acceptable standard of privacy and security, there needs to be a much more collaborative approach.

Because our industry is so expertise-heavy, it’s difficult for external innovators to build in this space without that experience. This is one reason why so many of the outside firms are building products with only peripheral knowledge of the issues that lawyers face. But being locked down with one company and not being able to sell elsewhere is a disincentive for those innovators.

In order to build developments that can benefit the industry more widely, collaboration between firms and companies is going to be important. This collaborative competition economical approach isn’t a new idea - it’s been used in everything from biochemical research to automobiles.

In the long-run, we’re going to continue to need innovation in this industry. If we’re led by fears around AI to the extent that a third have no plans to invest, then those companies just aren’t going to be part of the evolution - or by the time they get on board, they’ll already be behind.

‍

For more information about people’s responses to our survey, check out our report - The State of LegalTech Adoption.

At Definely, our starting point was the problems that lawyers face on a daily basis. We approach our own AI tools from that same starting point, and that includes understanding the security and accuracy challenges that firms and legal departments face. If you'd like to find out more about our tools and approach, please do check out the Definely Suite.