Loyola University Chicago School of Law, JD 2020
Artificial intelligence is all around us. Whether it exists in your iPhone as “Siri” or in complex machines that are detecting diabetic retinopathy, it is constantly growing and becoming a regular part of the modern day. As with any new technology, regulation surrounding artificial intelligence is becoming increasingly problematic. The question facing us now is how do we encourage further development without accidentally hindering its growth? Recently, the Food and Drug Administration has attempted to take steps toward further regulation of artificial intelligence by introducing a review process for medical artificial intelligence. This is just one instance of how regulation may affect the evolution of artificial intelligence.
What is Artificial Intelligence?
Artificial intelligence is the science behind creating intelligent machines and programs that can apply principles of human intelligence. It has led to “sous chef” Chef Watson, the monitoring of vehicle performance by Volvo, and even assisting in combating counterfeit Burberry products in the fashion world.
Artificial intelligence is categorized into two types: (1) narrow artificial intelligence and (2) general artificial intelligence. The former is the one we are more familiar with – virtual assistants like Siri and vision-recognition systems on self-driving cars. General artificial intelligence is more what we associate with robots in science fiction films – it is an adaptable intellect that is found in humans. Regardless of the type of artificial intelligence, the challenge of how to regulate it the largest obstacle.
Is Regulating Artificial Intelligence Even Possible? If Yes, How?
Some argue that regulating artificial intelligence may be impossible, stating that we do not have the proper ethical, legal, and governmental ability to monitor this growing area. While others ask whether artificial intelligence should be regulated at all, stating that regulations could hold back the exponential growth of progress.
However, growth of artificial intelligence and research surrounding it has shown the harms that can arise with a lack of regulation. Whether it be with concerns surrounding bias, with evidence showing that artificial intelligence programs can discriminate against people of color, to concerns around artificial intelligence being used for surveillance by the government. For these reasons, among others, proper regulation is essential to avoid its misuse.
Challenges extend to all areas surrounding artificial intelligence. Suggestions of holding individuals responsible for the actions done by artificial intelligence have been raised where it has been suggested that responsibility will be allocated to the creator or company in charge of said “intelligence.” Another interesting proposal rests in focusing regulation not on the research of artificial intelligence, but rather on the applications of it within areas it impacts – whether it be transportation, medicine, politics, or entertainment. This too is not foolproof, but could lend for a starting point on how to regulate artificial intelligence.
The FDA’s Step Towards Regulation of Medical Artificial Intelligence
The Food and Drug Administration (FDA) has already approved the use of artificial intelligence for medical devices that do not change the algorithm each time it is used. However, it proves to be more difficult to regulate a product whose algorithm constantly changes with each use.
Accordingly, the FDA recently set forth a framework to assess and regulate artificial intelligence products that are used in medicine. In short, the report states that the FDA will require a proper review before any medical product that relies on artificial intelligence is able to be commercialized. More specifically, it allows the FDA to review: the performance of the algorithms, any plans the manufacturer may have to modify anything, and the manufacturer’s ability to manage risks that may come from any modifications.
The FDA’s next challenge will be to find the perfect balance between (1) regulations and (2) improving medical care artificial intelligence machines without hindering the former.
Looking Forward & The Next Steps for Artificial Intelligence
Aside from artificial intelligence in medical devices, regulations for other sectors have yet to be established. The challenge lies in whether artificial intelligence, as it evolves, becomes aligned with human intelligence and thus should be regulated in the same way. This is a problem for the future, when—and if—artificial intelligence extends to such a level. For now, the focus should rest upon ensuring that progress is not impeded upon by regulations.