Global AI and Data Science

 View Only

Artificial Intelligence and the Role of Expert Witnesses in AI Litigation

By Anonymous User posted Tue March 16, 2021 05:33 PM


AI is in robotic systems, software, automated administration systems, mechanical operations and computer networks that control large portions of our techno-cultural and financial lives.

In a lawsuit involving AI technology, laypersons can find it very hard to determine the responsible party and the proper interpretation.

It is essential to enlist an Artificial Intelligence internet expert witness who can interpret the technical specifications for the lay public and bring the core data to light.

Thinking that AI can be found in so many kinds of computer systems and applications, it is important to hire an expert who specializes in the area of study particular to the case.

Artificial intelligence (AI) and machine learning (ML) have transformed the ways in which computers, users, and their computing environments communicate. In doing so, AI technology has started up a vast new horizon of possibilities to innovate and change nearly every area. But it has also opened up further issues for litigation!

What is Artificial Intelligence?

Artificial intelligence (AI) is a general category, including any software or device that can understand its environment and adapt its operations to reach its goals. The term “artificial intelligence” is usually used as shorthand for a machine doing stuff we usually connect with human cognition, like learning and problem solving.

The study of AI was formalized as academic training in 1956. The types of technology typically considered Artificial Intelligence tend to increase over time, often in relation to what people perceive as novel or routine. 

Major Problems in Litigation Involving Artificial Intelligence

Over the past several decades, everything from automobiles to household devices have become more complex. In many situations, AI only adds to that complexity. For end-users of AI results, determining what went wrong and whose negligence is to blame can be even more complicated.

One of the most significant challenges among AI problems is that artificial intelligence methods are not “programmed” in a general sense. Traditional programming includes giving an exact and detailed set of guidance to a machine, which brings out those instructions and displays the outcomes. Conversely, AI systems are designed to intelligently and dynamically change their instructions as required to suit changing input or environmental circumstances.

Knowing how the system changes is often crucial for fact-finders, but it can also be challenging to describe in terms both simple and straightforward enough for non-experts to understand. Here, a specialist witness who knows AI and explains it lucidly may play an indispensable role in the matter.

Authorities and Their Role in AI Cases

AI can be performed in a number of ways, using different sorts of computer instruments, methods, and modeling. As a decision, choosing the right specialist in an AI-related case will depend heavily on the particular experiences of the case and the most likely examples of what went wrong and why.

Experts engaged in AI cases typically arise from fields like computer or mechanical engineering, data systems, data analysis, robotics, and programming. They may practice in areas such as tools, software, 3D-printing, biomechanics, Bayesian logic, data science, e-commerce, or other methods.

Future research of artificially intelligent systems' legal situations may further confuse these cases and demand a more prominent need for experts. For example, the European Commission lately considered whether to give legal status to specific robots. One of the results weighed in the choice involved legal liability: Who is to blame if an AI-based robot or system, working autonomously, causes harm?

The European Commission failed to give robots legal personhood but did approve giving them a legal-entity status comparable to that of corporations. As technology changes, however, it’s not hard to imagine a near future in which the help of an expert in theory or ethics would also be necessary to resolve questions of fault in an AI case.

And while AI can make some cases more difficult, it can make litigation as a profession easier. Research into the use of AI for interpreting patent claims, performing record review, and detecting certain types of fraud is continuing, and it may produce benefits for attorneys and customers in the future.

Machine learning practices in daily use

Support systems from Amazon, Netflix and Google are examples of machine learning. Rather than people writing a program for a computer, a ML system takes a large collection of historical data and uses it to create a relation among an input and an output.

In a movie recommendation system, the data are the movies you watched recently, and the output is a list that the system “thinks” you might like. People are often amazed by how exactly these methods predict content they will also like.

Medical imaging is another pattern. Here a ML system has analyzed the different types of tissue in the breast, based upon a library of models previously diagnosed by physicians and used by the ML system.

If the information is a medical image, the output could be a tissue type and a diagnosis. Provide the ML system with enough pictures and past diagnoses, and one can get a system that produces “diagnoses” from new pictures.

These apps quickly become controversial. Radiologists have years of practice in image interpretation, and their expert knowledge delivers key insights for patient care. Assigning diagnoses and patient management choices to an AI network would raise a host of questions.

Dramatic ML progress at Google

In November 2016, Google Translate suddenly started to produce usually indistinguishable translations from the output of human translators. Machine learning has made slow improvements from its genesis in 1956, but until last November 2016, humans were still better at translating. Google's change was to use a neural network rather than their past phrase-based statistical machine translation method. 

How does ML affect litigation?

Litigation is a difficult and overwhelmingly personal effort. How might ML influence the process? Here are some sections where machine learning innovation is taking place.

One avenue of investigation is in the textual study of patent claims to predict a PTAB or court challenge's probability of winning. Having a score to describe the likelihood of winning a case, based on the IP landscape and applications, would have high value to businesses and attorneys.

Litigation processes could also profit from machine learning methods. Traditional litigation means combing through large quantities of documents and email in a search for connection. Keyword searching has been possible for years, but this would go well ahead in reducing the number of records that humans need to analyze.

Health insurance scams are difficult to identify. Recent ML research creates a numerical number for each transaction, and even a score of the likelihood that the transaction is fraudulent.

Machine learning is expanding its use into almost every field. For lawyers dealing with matters including machine learning, a machine learning/artificial intelligence expert witness actively operating in this area can give invaluable advice. Considering the use of AI/ML in intellectual property analysis, an ML/AI expert will certainly be required in patent and intellectual property disputes as well.