Home Healthcare Demystifying 3 Common Myths about AI in Drug Discovery

Demystifying 3 Common Myths about AI in Drug Discovery

Hype tends to be either a curse or a blessing, or both, especially in startups. Although hype about a market, technology, or a company opens doors, it also tends to trigger over optimism among believers, confusion among misinformed public, and skepticism among doubters. Such scenarios are being witnessed with AI in healthcare and this has led to emergence of some myths about AI and drugs. Read on to learn more about these myths.

It Is Easy To Develop
Some people argue that it is easy to use artificial intelligence in drug discovery. Such people suggest that widely available data and technology is already being used in medicine, and there are numerous machine learning researches and projects that are open source. Although this is all factual, the problem lies within the details. Currently, machine learning demands:

i. Precisely labeled data
ii. Including engineering for the purpose of assessing the most important features in the data
iii. A process for building, updating and cleaning datasets
iv. A user interface for the end user to access results
v. Ensuring results are always accurate by maintaining quality assurance

Most of these details require expertise. While researchers are exploring newer machine learning techniques and server options, the biggest debate involves accuracy, utility, and comprehensiveness of data and trivial details of the elements of a user interface. This shows that details matter a lot.

It Will Do Everything for You
There is a myth going around that AI can fully automate tasks involving drug discovery. Although AI will automate or disrupt many professions to increase productivity by augmenting human labor, it cannot fully replace people. No AI system can develop drugs without involving a human.

For instance, BenchSci uses machine learning for reading, extracting, and organizing unstructured data about antibodies studied in scientific papers. However, scientists still have to filter the results and use their judgment and experience to select the ideal antibody for specific work. Hence, AI is empowering humans and computers to do faster and better work while increasing productivity and reducing the costs.

It Is Inaccurate
Another myth going around is that AI is unreliable due to having inaccurate predictions. This myth is as a result of AI failing to be 100% successful on complex tasks. However, humans are not perfect either. They are often irrational and inaccurate but have still achieved a lot. Hence, AI researchers generally benchmark the performance of AI systems against mimicking the human brain and not perfection. For instance, the error rate of human transcribers is almost equivalent to that of an AI system. Humans have been found to have a transcribing error rate of 5.9% while that of machines such as Siri in 5.1%. Therefore, the accuracy rate is 94.1% 94.9% respectively. This shows that AI systems are actually more accurate than humans. Machine learning technologies are now extracting

Although machine learning technologies are assisting in extracting and labeling data, humans are always assigned to role of correcting any mistakes. The best data should be accurate, effective, and reliable. There might be more myths out there, and they all need detail assessment to indicate whether they are factual or not.

Source Bench Sci

Subscribe to our newsletter

Signup today for free and be the first to get notified on the latest news and insights on artificial intelligence

KC Cheung
KC Cheung has over 18 years experience in the technology industry including media, payments, and software and has a keen interest in artificial intelligence, machine learning, deep learning, neural networks and its applications in business. Over the years he has worked with some of the leading technology companies, building and growing dynamic teams in a fast moving international environment.
- Advertisment -

MOST POPULAR

AI Model Development isn’t the End; it’s the Beginning

AI model development isn’t the end; it’s the beginning. Like children, successful models need continuous nurturing and monitoring throughout their lifecycle. Parenting is exhilarating and, if...