How to use AI tools effectively
Hin Han Shum of Squire Patton Boggs
Apr 14, 2020
Jump to section:
1. When using AI tools what do we need to consider?
2. If we use an AI programmes for advisory services - who has the liability?
Video transcript
1. When using AI tools what do we need to consider?
There's a really strong appeal to using AI tools. When we think about AI tools, we think about how intelligent it might be, how it's able to predict things, how it can quickly grasp certain types of information that are required in order to streamline processes. But ultimately AI tools actually depend on what type of data is fed into it. So when using AI tools, it's important to consider, firstly, what the AI tool actually does. If it is in relation to sorting or categorising information or people, it's important to find out what type of big data has to be inputted into the system in order for it to ensure that it produces a result which is aligned with your company's approach at dealing with this type of information. This would actually be quite a big process to ensure that the proper and relevant documents, materials, are provided to the AI system so that it can grow its understanding and develop. This is one consideration, the data bias. Another important thing to consider in relation to AI tools is how it is implemented and how it is executed in a company. If it is, for example, an application that has an AI program running or an online mobile or a website application, it's important to inform the users of the applications, as to what type of data is being collected, how it's processed, and where it's going to be stored. These are all to comply with data privacy regimes in order to protect data privacy of the data subjects.
2. If we use an AI programmes for advisory services - who has the liability?
When it comes to AI tools and where the liability lies, there's actually several parties involved. It could be the programmer of the AI program. It could be the operator of the AI program. Or it could be the person who is running point on the program. For example, it might be the one who actually presses a button in order to ensure that the program runs. It might be the one who is maintaining the systems. These are all parties who are potentially liable when it comes to a problem arising with the program. But, when companies implement such AI programs for their business operations, customers who deal with the company would rely on the company. So, any company who wants to implement and use AI programs needs to make sure that they actually know what they're implementing in their system. They need to understand what are the risks involved. They need to understand how the program is run. They need to understand what type of problems can arise. For example, what type of tortious problems? Negligence? If it is an advice, is the advice going to be accurate? How do they ensure that? How do they ensure that it is delivered properly? How do they ensure that it is up to standard? These are all things that have to be considered when running these programs. And so, it's important at the onset to ensure that there is proper due diligence conducted on those programs, and also a system in place in order to monitor it, to make sure that it is kept up to date, that there are no problems, that it is, as to the best extent, that there is no technological problems, for example, viruses, potential loss of data, storage issues, destruction issues of data. These are all things to take note of.
More videos from Hin Han Shum, Squire Patton Boggs
Latest Articles
Latest News