SAA Coffee Break: AI and Connectivity in Advanced Manufacturing – Legal Implications

Many advanced manufacturing companies are seeking to integrate, or expand their use of, AI, connectivity, and other related technologies. While such tools can have significant benefits, they also bring potential risks and pitfalls. This coffee break will provide an overview of the key legal implications and potential risks associated with use of these cutting edge technologies in manufacturing for the automotive industry.

Missed the live session? Check out the recording and transcript below:

Transcription

The below episode transcript has been edited for clarity.

Ann Marie Uetz

Thanks everyone for joining us. I’m Ann Marie Uetz from Foley & Lardner. where I’m a partner in our Detroit office focused on manufacturing.

This is our first coffee break with the Society of Automotive Analysts (SAA) and we are really happy to be the exclusive legal partner with SAA. I sit on its board, it’s been a great relationship for the last couple of years. Foley is a full service law firm with 25 offices worldwide and 1,200 plus lawyers, and one of the four sectors that we focus on is manufacturing. So I like to say, our sweet spot is manufacturing and specifically automotive.

Today we’re going to discuss some AI issues that are timely and related to connectivity and advanced manufacturing, while talking about it from a legal perspective.

I’d like to introduce my partner, Nick Ellis. Nick is in our Detroit office and is all things supply chain. He works up and down the automotive chain. We represent a lot of suppliers and work in the electric space as well. So Nick is going to join us, and we’re happy to be joined by our partner, Shabbi Kahn, who also works in manufacturing and is really connected with these AI issues, as you will hear from him as he opens up his remarks.

I would really invite your feedback. It’s a 15 minute quick hit and we’re going to try to give you some informative points today, and you can always follow up with us.

We really welcome these so thank you to Kevin and everybody at SAA for setting this up. And with that Shabbi I’m going to hand it over to you.

Shabbi Khan

Thank you everybody for attending.

This is 15 minutes so I’m going to try to talk fast so we can get as much information to you all as we can. I think we’re going to go over three big topics. I’ll talk about ownership of data and other intellectual property, and then Nick will cover the other two. But to the extent people have questions feel free to use a chat to type in whatever you want. We’ll try to get to them, otherwise reach out to us after this and we’re happy to chat more.

So just talking about AI in general, I think people have heard data is the new oil as a cliche. I think it still holds true. And with AI it’s becoming more and more important to understand the relationship of data when talking about partnerships with AI vendors.

Assuming you’re a manufacturing plant or a facility, you might be trying to integrate AI into your workforce. That might mean partnering with an AI vendor, AI solution, and so forth. In a traditional relationship, the manufacturing plant will give access to certain amounts of data to these AI vendors, which can then process and parse that data into providing meaningful insights and action items that can help improve the performance of your manufacturing plant.

This could extend beyond manufacturing, as you can imagine, into telematics for automotive driving or autonomous driving and so forth, but keeping true to the theme of manufacturing let’s just assume that it’s typically an AI vendor that’s trying to partner with a facility to deploy AI.

What are some of the issues that pop up that I think companies or manufacturing plans should be thinking about? Obviously, ownership of data. And I say that broadly, because it’s not just ownership, but it’s also the data usage rights that extends from it. So you know, you may enter into an agreement saying that you own the data. But the AI vendor may obtain certain rights to use that data in certain ways. Or I think things get tricky in how much you know, how many rights you’re giving away with respect to that data. And what do you consider to be proprietary for yourself?

For AI vendors, their goal is to collect data from a whole bunch of different manufacturing facilities. You know, not just for you, but from others, and then sort of try to aggregate that data and compound the value of the aggregated data.

What that means for a manufacturing facility is you lose your bargaining power when you try to license that data separately. So just being aware of the issues around who owns what rights and what rights you’re giving away is important.

AI vendors, for example, in their in the agreements that you’ll typically enter with them, will have some language that says “we can use your data to improve our products and services” and that’s a pretty broad definition, because they can define products and services extremely broadly to even cover cases that they haven’t even thought about at the time they signed the agreement.

With generative AI, a lot of companies are announcing “oh, we already got the rights to use their data to improve our products and services, which means we can use their data to train our generative AI models or use it for other issues.” So just appreciating the nuances around data ownership, the rights to use and restrictions around it. One of the important restrictions that you always want to think about is when you terminate a relationship with a vendor. You want to make sure that they delete that data so that you retain ownership of it. Sometimes they won’t agree to that. But that’s a sticky point that warrants a further discussion around how you want to handle it.

I’ll quickly touch upon other intellectual property issues. Obviously, when you’re starting to work with an AI vendor, you know you most likely will come to the table and say “how can you help us?” and there’ll be some collaboration. We advise clients on both sides to actually come prepared with some patent filings in advance, so that you’ve already marked the territory of where you think your innovations may lie. Sometimes it’s “hey use this data to improve.” You know the performance of some material or a fastening technique of screws. And if you’ve already got those ideas in place, you might want to consider filing.

Last thing I’ll say about IP is with generative AI, there are issues that are creeping up around inventorship and authorship. Output by generative AI, in the United States at least, typically is not copyright protectable unless a user or an author has provided significant human contribution to the prompt and worked through the generative AI model. Without copyright protection, what that means is you’ve got limited rights in terms of other people being able to use the same designs, or whatever materials that you’ve generated through generative AI, for their own benefit.

And then from an inventorship standpoint, if you ask ChatGPT “how would you solve this problem” or “give me solutions” you start running the risk of not being an inventor, and therefore those ideas may not be protectable. They are carveouts that I’m happy to chat more about, but with seven minutes left I want to turn this over.

Nicholas Ellis

Thank you. Shabbi.

I’m going to talk a little bit about everybody’s favorite topic in the world, which is liability. And how do you apportion the cost between the parties when things go wrong? I know as we’re getting into these new relationships with new technologies there’s at least the hope that everything is going to go smoothly, and we’re not going to have any significant bumps in the road.

But unfortunately, as we’ve all experienced in the past, something is probably going to happen at some point, and the hope is just that it’s nothing major. But if it is, you want to have a good division of responsibility set up in the contract so you’re not having to fight about that later.

I think we’re currently seeing these AI and connectivity issues coming to play. A lot of it is in the manufacturing systems and the equipment itself. As we all know, the biggest risk is an interruption in production which can result in significant liability. And so when that occurs, you have to think about who ultimately is going to have to pay the cost for that. Do I have recourse against my vendors? Either the provider, the equipment, the provider of the software, etc. Where does that responsibility lie?

And you know, the tricky part in a lot of cases is going to be figuring out who has the responsibility for the integration. So when there’s a production issue, is that something from the software or is that something that flows from the equipment. And do you have one point of responsibility for those issues?

There’s really no great magic to how you can deal with that. It comes down to the basic blocking and tackling of your contracts and having good agreements in place. You want to understand what the warranties are, what the indemnity provisions are, what the responsibilities of the parties are, and then critically, what are any limitations of liability that you may have had to accept as part of those contracts.

I can have the best warranties and the best indemnity provisions in the world. But if there’s a limitation of liability that says the most I can recover is the purchase price, and now I’m staring down the barrel of a 10 million dollar charge from my OEM customer for a prolonged shutdown, that’s not going to get me very far.

So you have to understand what it is you’re signing up for. What are the rights and resources that you have. Which leads a little bit into the third topic here, which is the risks that are inherent in a lot of these new relationships that many suppliers are having to get into as they start to explore these new technologies and these new approaches to manufacturing a lot of the time. You’re not going to be dealing with your traditional automotive customers.

And you know, a lot has been made about the clash of cultures between the automotive industry and Silicon Valley. And  in some cases that may be overblown. I think a lot of people outside the automotive industry, they look at the contracts in terms of conditions that are standard in the automotive industry, and we all take for granted. And those can generate some pushback at times when you’re dealing with these new technologies, especially AI systems – a lot of the time you find yourself dealing with extreme ends of the spectrum.

As far as your partners, you may be dealing with a Microsoft or a Google; somebody who is the proverbial 800-pound gorilla in their own area. And you’re going to have very limited ability to push back and negotiate on the terms. You may very well be in a situation where you just have to either accept the terms that they are offering or you’re not going to be able to use that technology.

And you know, those types of contracts are likely going to be very favorable to the vendor and limit your recourse on the other end of the spectrum. You may be dealing with a company that is more in the realm of a startup, and while you may have great leverage to try and negotiate contract terms with those types of vendors, there’s going to be concerns about whether the vendor can really stand behind those obligations both from a performance standpoint or a liability standpoint. Again, I can have the best indemnity provision in the world with no limitation of liability. But if the other side doesn’t have the assets to support when I have to make good for a 10 million dollar shutdown charge, you know that contract is not going to be worth the paper it’s printed on.

And so that’s a risk that you have to deal with. Its really just understanding who are you getting into business with? What is this company? What’s its capabilities? What’s its assets? Can it really stand behind what it’s committing to?

Shabbi Khan

Nick, I was just going to say the other thing to think about when you’re dealing with smaller or newer companies is that trust becomes a big factor. How much data you’re willing to provide to them and can you trust that they’re going to keep it secure. When you’re thinking about a Microsoft or Google, you know that that data is going to be secure. And so sometimes cybersecurity breaches and confidentiality comes into play, depending on the nature of your entity as well.

Nicholas Ellis

Absolutely. So to close, while AI and connectivity have the potential for great benefits in manufacturing, both in terms of efficiency and output, they do come with risks. It is important to make sure that you understand who you are dealing with. Make sure you are complying with all the applicable laws and regulations surrounding data and privacy, and make sure that you have appropriate contractual protections in place and understand the risks that you may be left with.