Ethical Concerns About Using AI in Pharma

Artificial intelligence in pharma refers to the use of automated algorithms to perform medical tasks that traditionally rely on human intelligence.

While there are plenty of advantages to implementing the technology, decision-makers remain cautious.

According to Healthcare Weekly, executives across the pharma industry are looking at ways to leverage AI in their line of business, including healthcare (or the biotech industry to be more precise). 

Addressing ethical issues will enable the healthcare and pharma industries to move forward and potentially save millions of lives.

Here are the top ethical concerns about using AI in pharma:

• Data protection and privacy
• Fairness
• Transparency

This article will discuss each of these concerns and how to overcome them.

Data Protection And Privacy

Artificial intelligence and machine learning algorithms are subject to data protection and privacy concerns. Developers must pay attention to the ethical implications at each stage of data processing.

Given the fact that for machine learning to ‘work’, any company needs to process large sets of data, it is imperative that customers are consenting to the processing of their personal information.

Using personal data in the pharma space must also be in line with new laws and regulations surrounding consumer data privacy, such as the General Data Protection Regulation (GDPR) that has been enacted in Europe. GDPR applies to data to all European residents no matter where their data is processed, including the US.

Data that is obtained must have the appropriate authorization before it can be used, but determining if it has been allowed for processing is difficult. There are many factors that determine if authorization has been obtained, such as the type of data recorded and its purpose.

For example, data collected by wearable technology and smartphone apps must be stored securely. Patients should be informed that by using wearable devices, their personal data will be shared with third party suppliers.

Fairness

AI in pharma and machine learning algorithms must avoid using biases when determining whether or not to process or validate data.

Bias in medical data comes in two forms; cases where data sources do not reflect correct epidemiology with a particular demographic and when an algorithm has been trained on data sets that do not contain enough members of a particular demographic.

Algorithms that include these biases are not able to make ethically correct predictions and decisions – they have the potential to widen the gap to fair healthcare.

In order to avoid this scenario, medical practitioners must develop technologies that can recognize biases. There is also a responsibility on institutions such as review boards and ethics committees to check that the necessary standards are being met.

In addition to this, official guidance is required in order to maintain safe and efficient applications of artificial intelligence and machine learning.

Since AI algorithms can learn on its own, in real-time, it is not possible to test each update individually during clinical trials. As such, AI companies must develop standard procedures that can clearly document how algorithms evolve, how they are monitored and updated over time.

Transparency

As I argued before, one of the biggest ethical concerns around AI in pharma relates to what the exact data which cannot be directly accessed by developers. This poses a specifically critical conundrum to pharma companies.  

This makes it tough to trust that the AI programs can do what they are intended. Also, the disclosure of basic yet meaningful details about medical treatment to patients means that the doctors must grasp the fundamental inner workings of the devices they use.

In order for AI and machine learning to be ethical in the pharma space, developers must work with healthcare professionals and communicate the logic behind the decisions made by AI algorithms. They must have clinical validation and randomized clinical trials.

In the event that decisions are made automatically, risks associated with the procedure may determine how data is provided to patients about the use of AI and machine learning in relation to their care.

Healthcare providers must be able to get the balance right in order to function ethically. While AI and machine learning has the potential to help, there must still be human dialogue and interaction for the best care.

Conclusion

These are the top ethical concerns about using AI in pharma. If and when decision makers can address these issues and come up with solutions to overcome them, both patients and healthcare professionals can benefit from the advancements in technology.

There are many applications that the healthcare industry can use AI and machine learning for, such as leveraging smart devices via the Internet of Things (IoT) protocols and even going beyond standard clinical settings towards managing diseases and conditions.

As companies continue to invest in AI and machine learning in pharma, more and more products are expected to enter the marketplace.

But it’s vital that companies and payers address privacy, fairness and transparency issues to allow the use of AI in pharma to boom in the years to come.