myliethach-blog

[myliethach-blog]

Is an “Artificial Intelligence Industrial Revolution” coming? Or is it already here?

Artificial intelligence (AI) has an inconceivable potential to revolutionize our lives. Whether we are ready or not, artificial intelligence will change the world. Artificial intelligence is intelligence demonstrated by machines. When a machine mimics cognitive human functions such as learning and problem solving, the term artificial intelligence is used. Although there are many examples of artificial intelligence being used positively, there are still critics that are afraid that true AI will be a serious threat to humanity.

As a healthcare professional, I am excited about the potential AI has in healthcare. There are already healthcare facilities using AI to process large amounts of data to save time and money. New AI technology is also being used to detect and prevent diseases which will save so many lives. But, as a concerned citizen, I am worried that AI will have an overwhelming presence in our lives and threaten the labor force.

This blog consists of the evidence I found during my research on AI in healthcare and medicine. I invite you to explore the various views and uses of AI. I challenge you to decide if we should support an “Artificial Intelligence Industrial Revolution.”

[myliethach-blog]

Medial EarlySign is an AI start-up company conducted a new study showing how the combination of AI and HER data can provide early detection and treatment of kidney problems. This information can be used to help slow down or even prevent the progression to end stage kidney disease. The study revealed that using machine learning algorithms could identify 25% more diabetic patients at risk for kidney damage compared to clinical tools and human judgement. This use of AI in healthcare is changing healthcare drastically in a positive way. Ai is commonly used to manage cast amounts of data and simplify daily tasks. Now AI is evolving in healthcare where AI is used to find complex patterns in medical imaging and designing patient treatment plans.

[myliethach-blog]

Elon Musk is famous for being the founder and CEO of SpaceX. He is also the product architect of Tesla. Even though he’s famous for his futuristic gambles, he is hesitant to embrace the boom of artificial intelligence. Elon Musk is a big investor in DeepMind, a company known to be the world leader in artificial intelligence research. Musk reveals that he did not invest in the company because he supports AI, but rather, it gave him more visibility into the growing rate of AI. Musk’s alarming views on the dangers of AI went viral when he spoke at MIT stating that it was probably humanity’s “biggest existential threat.” He believes that there should be a national or international regulatory oversight to ensure “we don’t do something very foolish… with artificial intelligence, we are summoning the demon.” Due to his increasing concerns over AI being a threat to humans, Musk and Altman founded OpenAI, which is a billion-dollar nonprofit company, with the goal to work for safer artificial intelligence.

 Although there are many benefits to AI technology, Elon Musk is weary to welcome AI with open arms. He has gone as far as creating a billion-dollar company to monitor AI to ensure its safety. This should be a red flag to the public. Although AI and new technology can be exciting, like Musk, we should also accept AI with caution.

[myliethach-blog]

Google AI made a huge breakthrough by successfully predicting cardiovascular problems by analyzing images of the retina. This prediction is possible because Google AI uses machine learning to uncover the new ways to predict the problems. The article stresses that it is not imitating existing diagnostic tools.  The AI uses a deep learning algorithm on data from over 200,000 patients to predict cardiovascular risk factors from retinal images with high accuracy.

This is an exciting new diagnostic tool that positively impacts healthcare. Another prime example of how artificial intelligence can be used in medicine to save lives and prevent disease. Unlike other diagnostic tools, this AI technology can be performed without the use of any blood draws. Since cardiovascular disease is the leading cause of death for people of most ethnicities in the United States, these AI retinal scans can save many lives and allow the patients and healthcare providers to work together to prevent heart attacks and strokes.

[myliethach-blog]

The Department of Veterans Affairs (VA) announced that they will be partnering with DeepMind, an artificial intelligence company, to take a proactive approach in patient care. DeepMind’s machine learning platform will help the VA identify risk factors for deterioration in patient condition and predicting its onset. The partnership will start by uncovering risk predictors for the most common conditions during a hospital stay, such as acute kidney injury. Then the partnership will focus on other indicators of patient deterioration.

This is a great example of how artificial intelligence will make a positive impact in medicine. Medicine is not just about treating the illness the patient presents. A big part of medicine is prevention. AI would allow healthcare providers to identify risks early so they can take the appropriate measures to prevent the disease from occurring. DeepMind’s ability to predict patient’s deterioration would allow the VA to intervene earlier than what they could without AI. This early intervention could mean life or death for a patient. If more hospitals could partner with companies like DeepMind with the intention of taking a proactive approach to medicine, AI would save many lives.

[myliethach-blog]

Here's what Tech Billionaires have to say about artificial intelligence. Elon Musk says that AI is a risk to human civilization. Mark Zuckerberg says he's optimistic because AI will improve the quality of life. Bill Gates doesn't see a need for panic, but he doesn't agree with the people who says that there should be no worries to AI. Most would agree that AI holds great promise for the future, but there are others who fear the changes that AI will bring. Who do you agree with? Where do you stand?

[myliethach-blog]

How can AI still be dumber than a year old when AI can be used to predict diabetes with just a retinal scan? Well this article explains it. AI is used in programming techniques for pattern recognition and have become more effective than ever at playing games and performing speech recognition. For example, AI is exceptional at playing games like poker and chess. Chess programs now can play at a level best described as “superhuman.” But what is inaccurate is the claim that AI will soon be able to perform complex tasks that involve common sense such as customer support, selling products, and negotiating business deals. AI would not be able to immediately adapt to the changes during a negotiation deal. For example, AI would be successful with playing a game with pre-defined rules that never require common sense. But if unknowns were introduced, AI would not be able to play in real life which requires adjustment and common sense. Therefore, computers still can’t exhibit the common sense or general intelligence of a 5-year-old.

This presents a huge problem in healthcare if AI technology is used for patient care. Healthcare providers must be able to make adjustments to treatment plans if the patient is not responding to the original treatment. AI can create more problems if it is not capable of adapting to various unknowns that are commonly experienced in healthcare. It is a scary thought to think of AI being used in patient care if it can’t exhibit the common sense or general intelligence of a 5-year old.

[myliethach-blog]

It’s undeniable that artificial intelligence has the potential to accelerate medical research and better detect diseases. All the excitement over AI in healthcare overpowers some of the negative impacts AI can have. This article lists four major things we should be worried about:

1.     Phishing scams could get worse; criminals can send legitimate with malicious links that are more prevalent and effective thanks to AI technology.

2.     Criminals can use AI techniques to automate tasks like payment processing, helping them collect ransoms more quickly. The technology can help criminals target more people at once without personally communicating with them.

3.     Fake news and propaganda will only get worse, because AI led researchers to create realistic audio and videos of political figures designed to look and talk like their real-life counterparts.

4. AI could make weapons more destructive. This is the scariest one to me. AI could enable a person to cause widespread violence. For example, a self-flying drone will can detect a specific person’s face and then carry out the attack.

Although healthcare and medicine is not listed in these four major things we should worry about, the list is an eye-opener to the dangers of AI in our lives. Although there are undeniable evidence that AI can revolutionize healthcare, we must consider the negative impact it can have on other aspects before fully embracing AI.

[myliethach-blog]

In my earlier post about Elon Musk, I said that he is one of AI’s most vocal critics, openly telling MIT that AI is humanities biggest existential risk. That comment went viral in 2014. In 2005, he founded OpenAI, a non-profit company working for “safer” AI. In 2017 he said that the United Nations needed to act to prevent a killer robots arms race. Now the tech billionaire has quit the board of research of Open AI. Many may view his resignation as a loss to the AI community of critics.

If Elon Musk, who was known to be a huge critic of AI changes his position on AI, should the public reconsider also? Don’t worry AI critics, Musk says since Tesla became more focused on AI, he quit the bard to avoid any conflict of interest. Despite his designation, Musk says he will continue to donate and advise the group.

[myliethach-blog]

DeepMind technology made a breakthrough in 2013 when they wrote a paper explaining how a neural network could learn to play a video game the way humans do- by looking at the screen. Google then bought the company and has gone on to apply deep learning in a variety of situations to outperform humans. Although it is quite impressive, it also highlights the limitations of deep learning. Compared to humans, the machines using the technology take a longer amount of time to learn.

This is similar to an earlier article on how AI is dumber than a 5-year old because it couldn’t exhibit the common sense a 5-year old has. This article provides a great example of how AI will fail if unknowns were introduced intro the equation. Therefore, true AI is not possible yet.

[myliethach-blog]

The biggest argument that can be made against the use of AI in healthcare is the risk to patient privacy. The Health Insurance Portability and Accountability Act (HIPPA) establishes strict policies and procedures for maintaining patient privacy and the security of the individual’s health information. HIPPA also establishes civil and criminal penalties for violations to these policies and procedures. Since patient privacy is a major issue, all new technologies are adopted carefully to ensure that it is following privacy laws and protecting the consumer’s health information. This forces the healthcare industry to have a balancing act between privacy and innovation. The concern for privacy forces the healthcare industry to incorporate AI into their system a lot slower compared to other areas.

[myliethach-blog]

Another great use for AI in healthcare. Physicians often work long hours and often are on call which may not give them enough time to recover and recharge their batteries between shifts. This can lead to physician burnout. AI can help prevent physician burnout because they can recognize patterns in large quantities of data, saving he physician time and energy. For example, AI can identify matching images of X-rays for radiologists, saving them time. AI can also improve searching tasks and the documentation process, all of which contribute to physician burnout because it can be time consuming. Studies show that AI can help physicians reduce the time they have their hands on the keyboard.

If AI can be used for physician burnout, I would like to see how it can be used to help prevent nurse burnout. Nurses often leave bedside nursing because of burnout, which contributes to our nation’s nursing shortage. If AI can safely be incorporated in patient care, it can potentially prevent burnout and help with the shortage.

[myliethach-blog]

Accenture analysis says healthcare’s artificial intelligence market will reach to a staggering $6.6 billion by 2021. With technological boom, it is expected to see an increase use of digital technologies to control when, where and how patients receive health care services. As many as 1 in 5 consumers say that have already used healthcare services powered by AI through virtual clinicians and home-based diagnostics. This makes me worry about the quality of care patients will receive. Yes, a great volume of care will be delivered but quality of care may be compromised.

[myliethach-blog]

The previous article discussed how AI can be used to help physicians by preventing physician burnout, but this article presents a great point. If AI can perform some of the physician’s tasks, wouldn’t many physicians be out of a job? A renowned Silicon Valley investor claimed that 80 per cent of future doctors would be replaced by AI algorithms. I think that is an exaggerated prediction, but there is some merit to his claim. Currently, AI cannot perform on the human level of intelligence, but at the most basic level, AI can conduct many of the tasks that we can perform. This is where AI can make future physicians out of a job. For example, there is evidence supporting the use of algorithmic approaches to identify tuberculosis patterns on chest x-rays and like a previous article I include earlier, screening imagery can isolate early signs of diabetic retinopathy.

[myliethach-blog]

It is easy to hype up artificial intelligence when it has proven to be beneficial in healthcare. The biggest fear of AI is that the machines will replace or even turn on humans. It is true that computer intelligence is advancing faster than human intelligence. If the computer speeds double another 5 times for the next 10 years, machine learning tools and diagnostic software could become as essential to healthcare providers as the stethoscope was in the past. Although AI can save time by improving efficiency through processing data at a faster create, computers have yet to demonstrate the kind of empathy and compassion that millions of patients rely on.

[myliethach-blog]

There is no doubt that the rise of artificial intelligence has the potential to transform healthcare by improving productivity and patient outcomes. But Elizabeth Clements, a business architect, says that many health systems do not have a business plan in place to support the advanced analytics beyond research and development. As I think about the effects of AI in medicine and healthcare, I often worry about providing quality care and protecting patient privacy. I never thought about the business side of it. For healthcare systems to have success with AI, they must consider the business implications. Clements suggests a simple framework for thinking about machine learning in the context of business. With a good business framework, implementing AI in the healthcare system would yield a higher success rate.