Episode 12 – Ethics and AI within the Health Industry

FortressIQ | Intelligent Insights for the Modern Enterprise
Episode 12 - Ethics and AI within the Health Industry
/

The next generation of AI-powered products and services has the potential to transform every aspect of the health sector. From accelerated drug discovery through biophysical modeling to improved diagnostic accuracy in medical imaging to smart claims management with self-learning software, no corner of the $10 trillion global health industry will remain untouched. Someday consumer-facing apps may become so advanced that they are able to provide affordable healthcare services to billions of people globally with limited access to traditional services. But these advances do not come without risks, as biases in AI raise ethical concerns and present health risks to patients and financial risks to providers and payers.

Today’s guest is the brilliant Zahra Timsah. Zahra has a multitude of professional and academic achievements including a Masters in MBA, and PhDs in Advanced Molecular Biology and Cancer Biology. Zahra is the lead for AI Governance at MassMutual, the Co-Founder and COO/CTO of Assurance Medication Management, Inc, as well as the CEO of AMCL.

In today’s episode, host Jon Knisley, long-time technologist helping companies win the market with emerging AI technologies, and series producer Elizabeth Mitelman talk with Zahra about the importance of AI in the healthcare industry. They also touch on the ethical risks in having AI advancement in this field, and the best practices to assure that the proper AI governance is in place.

Talking Points:

  • Highlights in Zahra’s professional and academic career
  • Zahra’s experience integrated AI into the drug discovery process
  • Other opportunities in the medical field where AI can solve other challenges
  • How Zahra turned her fascination with AI into a career
  • Advice for people who have a passion, but are unsure as to whether or not the passion will lead to a successful career
  • Best practices for ethics and AI governance
  • Zahra’s role in Mass Mutual
  • Advice for the next generation for women in AI

Resources/Links:

Zahra Timsah (MBA, PhD, MSc, BSc),LinkedIn

FortressIQ

If you enjoyed this episode, subscribe and check out our series atfortressiq.com/podcast. Thanks for joining us today on hello, Human.

Full Episode Transcript:

Jon: Zahra Timsah, the AI Governance and Ethics Leader of MassMutual Life Insurance, joined us today on the hello, Human podcast, where we discuss the latest topics in artificial intelligence and how it’s being applied in the real world. I’m Jon Knisley, the host of hello, Human and a long-time technologist helping companies adopt and utilize emerging digital tools.

A big thanks to FortressIQ for sponsoring the program, and be sure to subscribe wherever you listen to podcasts. This episode is part of a special series on women in AI that we are very excited about here at FortressIQ, and Elizabeth Mitelman from our marketing team who’s been driving the program and is participating in the session as well.

In this episode, we’re going to explore ethics in AI within the health industry. AI-enabled tools have been used in pockets of the sector for some time now, especially drug design models and medical imaging. But the next generation of technology promises even more, and no corner of the $10 trillion global health industry will remain untouched. 

Someday, consumer-facing apps may become so advanced that they are able to provide affordable health care services to billions of people globally with limited access to traditional services. But these advances do not come without risk as biases in AI raise ethical concerns and present health risks to patients and financial risk to providers and payers. It’s a fascinating topic and we’ve got an incredible person to discuss it with. 

Welcome to the program, Zahra. Thanks again for joining us on hello, Human and participating in our Women in AI series. When we were first introduced, I had to admit that I was intimidated by your credentials. You’ve got all sorts of acronyms after your name. You hold a Master’s, an MBA, as well as a PhD, and not in fluffy topics. We’re talking advanced molecular biology, as well as cancer biology from MD Anderson—one of the preeminent research institutions in the world. 

On top of that, you served as CEO of an AI and big data company and CTO for a health care and insurtech enterprise. And now, you are leading AI governance for a Fortune 100 company. Just an incredibly inspiring background. Can you maybe highlight one or two of your proudest achievements in your educational and professional career?

Zahra: Hi, Jon. Hi, Elizabeth. So glad to be here today, and thank you guys for inviting me. I think your team is doing wonderful work. It’s highly educational and it’s highly insightful, so thank you. 

Let me just give you a very quick overview about myself, and then I’ll highlight a couple of basically my proudest achievements, so to speak. I started my career early on in my life. After getting my Bachelor’s and Master’s degrees in Molecular Biology and Biochemistry, I worked for a pharmaceutical company at a very young age to create medications for prostate cancer. The drugs were actually approved and went to market, so that was amazing. 

Back then, I discovered that I have a passion for biostatistics. I appreciated the importance of the development and application of statistical methods in the design of experiments, the collection and analysis of data, and results in its implementations. Early on, I also discovered that coding and programming languages can help advance the health care and medical field, so I started researching, learning more and more about machine learning and AI.

One of the most exciting periods in my life was actually my PhD and postdoctoral training time at UTHealth MD Anderson here in Houston. My focus was cancer, drug discovery, and development. I was lucky enough to publish in high-impact journals. Shortly after—because of those achievements—I got selected to be part of 250 Great Minds, which is a competition in the UK, specifically at the University of Leeds. I was selected as a tenure track faculty member. I was a principal investigator with my own team and centralized funding. I also landed Wellcome Trust and Well Society funding. 

Then I decided to pursue an MBA while working at the university because of another critical observation. Business practices drive the healthcare industry, so I needed to know business, I needed to know management, I needed to know finance to ensure that valuable work is socialized and implemented in an efficient manner. 

Then I just decided to apply what I learned and worked as a consultant for companies like Bristol Myers Squibb and Johnson & Johnson during the era where cancer immunotherapy discovery and development was the buzz.

You know what they say—necessity is the mother of invention. With the life science landscape being increasingly complicated, I found out that companies must constantly evolve. What better way than the adoption of artificial intelligence, which we can discuss in a bit.

Another great achievement is when I got offered a job as the VP of Operations and the Director of AI at an immunotherapy company here in Houston. I help them develop a healthcare AI program from scratch. This advanced their R&D. It advanced their clinical projects as well. Partly because of that, they became publicly traded; they went IPO.

I would just end it by saying, the biggest achievement for me basically is starting my own company, starting my own startup companies. One of them is called Assurance and the other one is called AMCL. At both Assurance and AMCL, I work as an executive, as a CEO, and a CTO. AMCL is actually an award-winning AI and data analytics company. It has its international presence and it specializes in AI software and data-focused services. 

In a very short period of time, we managed to cater to financial institutes, to insurance companies, to healthcare companies, and so on. The company went from just being a startup and in a couple of years as revenue-generating. It’s producing millions and millions every year. You see, it’s all about generally caring, about advancing the field, finding a market gap, and jumping on the opportunity to be creative in an unaddressed market group.

Jon: That’s great, Zahra. Just an incredible background. Drug discovery notoriously moves at a glacial pace. I spent a number of years accelerating clinical trial startup and working in discovery, as well as helping automate the analysis of electrocardiograms at one point. As you recognized, it’s really a big data problem that can benefit from AI. Can you talk a bit about your experience integrating AI into the drug discovery process? Are there any other opportunities within the medical field where AI can solve other challenges?

Zahra: Sure, and I absolutely agree with what you said. Artificial intelligence is an overarching term that is widely used in the medical field. I can only speak from my experience based on the systems I’ve developed so far. I’ve used algorithms to try and mimic human logic and cognition so that we can analyze, we can understand, and we can present solutions to complex medical problems based on the training data set, based on the input data.

When it comes to drug development, the vast chemical space fosters the development of a large number of drug molecules. There is a lack of advanced technologies and this can limit the drug development process, making it a time-consuming and expensive task. This can be addressed by using artificial intelligence.

Artificial intelligence in the drug discovery field can recognize hit and lead compounds. It can provide a quicker validation of the drug target and, in a way, the optimization of the drug structure design. Artificial intelligence can really be used effectively in different parts of drug discovery. You can use it in drug design, you can use it in chemical synthesis and drug screening, and also to address polypharmacology. And drug repurposing. Why not? AI can help identify new therapeutic uses for old or existing available drugs. 

But that’s not the only example where AI has been used in the medical field. AI can be used to diagnose diseases. One of the areas where AI and healthcare has shown the most promise is in the diagnostics field. As you know, for the majority of diseases, early diagnosis is one of the most critical factors and the ultimate outcome of patient care. 

For example, take cancer. AI deep learning algorithms will use these to shave down the time it takes to diagnose cancer patients. The way AI rapidly processes this large amount of medical information and arrives at likely medical causes for the cancer symptoms, can dramatically reduce the diagnosis-treatment-recovery cycle for cancer patients. The effects of this are already being felt and in several types of cancer—from leukemia, prostate cancer, ovarian cancer, and whatnot.

Another example that I can give for the use of AI in the medical field is radiology. Deep learning algorithms, AI algorithms can diagnose scan results so they can compare images with tens of thousands of other scans from healthy patients. For example, if you’re talking about head scans, they can identify potential lesions and abnormal regions in the brain. 

Let’s say, a neurological disorder like epilepsy. It spreads across the brain. If AI is used to identify abnormal scans as early as possible, this would be crucial in improving a patient’s treatment options and also, of course, the ultimate outcomes.

Elizabeth: Zahra, thank you again for joining us for the Women in AI series. We’re so happy to have you. In a previous conversation, you’re describing machine learning and AI as a hobby. How did your hobby transform into a career?

Zahra: Elizabeth, that’s an excellent question. You know what they say—choose a job you love and you will never have to work a day in your life. This is what I did. I found that there is a market gap and AI can be an essential part in driving businesses and innovation forward. After doing market research, I found the market opportunity where my skills and knowledge are unique and needed. 

My hobby transforming into a career wasn’t really driven by convenience but by prospect, by the need to feel fulfilled. I stayed true to my vision, I stayed true to my brand. I got enough exposure and I network a lot. I listen to feedback, I listen to criticism. Whenever you can, you should always listen to feedback and criticism because enthusiasm and passion can be a person-in-disguise. It really keeps us from seeing obstacles, especially if we’re trying to transform a hobby into a career. I made sure to factor in constructive feedback. I never stop brainstorming. 

My startups, the two companies that I started—AMCL and Assurance—deflect the hobby, transforming to a career. Both initiatives were driven by inventiveness, they were driven by imagination, but most importantly, the intention to encourage the market and meet the market needs.

Elizabeth: Do you have any advice to individuals who may have their own passions and interests, but they may be unsure if those interests could actually lead to a possible career pathway?

Zahra: Yes. I think it’s always important to ensure that the skills you have are unique skills; they are needed skills. It’s always important to research, it’s always important to read because, again, like I said it’s not just about doing something that is not just important for you to be skilled but it’s also important for you to fill a gap, to meet a need, so think about that. Is what I do unique enough? Is it needed? Is it addressing a problem? Can I do it better than others?

Jon: I’ve got to agree that when you love what you do it makes life much easier to get out of bed in the morning and get to work. Our CEO is very passionate about ethics in AI. Our official mission is to unlock the limitless potential of the global workforce by accelerating the responsible and ethical use of AI in the enterprise. It’s not surprising that we spend a lot of time and resources to make sure we have the right guard rails in place to avoid using our platform for any nefarious activities. 

You’ve got a background, obviously, working to ensure the proper governance for AIs in place. What about some best practices for an organization maybe just starting out in that journey, and maybe some best practices for companies that already have a program in place but want to take it to the next level?

Zahra: For organizations that have just started to think about AI governance, the case usually is to bring an expert onboard to ensure that any AI development efforts are in compliance with regulations and the principles of AI. When I say principles of AI, there’s actually a list of principles that I will explain in a sec. What these experts do is they’ll try to establish and implement morally responsible and defensible AI platforms from an ethical perspective. 

It’s hard to achieve, but AI principles must be translated into actionable steps so that they can guide and influence the company operations. For that, experts with both technical and business skills are needed most of the time, so that they can contextualize these principles into guidelines, into best practices that truly drive AI development and risk mitigation. 

Some principles that we need to focus on are the following: the first one being fairness. This principle is concerned with ensuring that the artificially intelligent systems don’t harm people and don’t harm customers through inequitable treatment. Biases in data must not be reproduced. They must not be accelerated by integration with artificial intelligence. This is something very important to keep in mind once you started to adopt an AI governance program.

You also need to ensure that the AIs being developed are developed with full trust and transparency. Many AI systems are black boxes. There is often a need for explainability, interpretability as well. This is another thing that you should keep in mind.

The third principle is accountability. As you know, there’s usually a result of a complex supply chain. You have internal and external elements. This involves data providers, data scientists, technology providers, and systems integrators. You need to have an accountability framework in place to nurture the sense of responsibility and ensure that AI systems are developed, again, responsibly and fairly.

Most importantly, the social benefit. Any AI governance program that’s being developed should keep in mind that AI systems must benefit the society. It must benefit people. If you look at a recent example, right now in our current climate this aspect of social benefit is crucial, especially for companies that are trying to leverage AI to develop a COVID vaccine. These companies are prioritizing human benefit over technological advancement because they want to achieve social benefit.

Another thing that one must keep in mind when creating an AI governance program is privacy and security. As AI systems are trained and then used to differentiate treatment, they need to respect an individual’s privacy. Between GDPR, between CCPA, this is the principle that has already seen the most legislation, so it is one of the most important aspects. 

Now, if AI governance program is already present at the company, in order to use it securely and efficiently, and a way of another mature it as an asset, a bond of assurance must be forged between developer and user, that a given AI outcome is being safe, it’s inclusive, it’s fair, secure, private, transparent, and accountable. This is when you bring an expert to ensure that AI governance functions are basically not static processes but processes that are dynamic, processes that are adaptive. Hence, the AI governance processes themselves must be dynamic and adaptive as well. 

Elizabeth: Those are all really great tips. Going back and looking at your career journey, you have had quite the journey in regards to your educational and professional career. Would you be able to share a bit more about your current role at MassMutual and your participation in the AMCL and Assurance advisory committees?

Zahra: Absolutely. First of all, MassMutual is one of the biggest Fortune 100 companies. It provides life insurance, protection products, investment, and retirement services, basically. At MassMutual, I specialize in leading the enterprise AI governance and ethics program, creating it from the ground up. It is something that I’ve helped other companies achieve as well. My current role can be summarized as leading the effort to ensure that the technologies, specifically AI technologies, are well-researched and well-developed in compliance with regulations, and with the goal of helping stakeholders navigate the adoption of AI systems further. 

I want to stress that the strategic and operational frameworks for AI governance revolve around the basic AI principles I previously mentioned. Basically, what I do is I make sure that there is a program that is a framework, a context in place to make sure that all of the AI systems being developed at MassMutual are developed in compliance with all of these AI principles, to ensure that they are monitored throughout their life cycle, and the risk is mitigated (obviously) relative to the risk appetite, relative to the risk tolerance of the company.

In my roles previously at AMCL and Assurance (and as an entrepreneur), I led not just the AI development and the governance but also fundraising, partnerships, collaborations, and obviously product development efforts. I had major operational, major strategy, leadership responsibilities. 

Let me point out that my job, in general, and stemming mainly from having expertise in both business and technology, is to bridge gaps between the two fields—between business and artificial intelligence—so that one can optimize success in the companies, to drive revenue generation, to drive customer satisfaction. Whether it’s at MassMutual, AMCL, or Assurance in the past, I create AI systems. I create AI governance programs that work, that are competitive, and satisfy market needs with state-of-the-art capabilities and in compliance with AI principles, with regulations, and ethical considerations.

Elizabeth: That’s amazing. My final question for you, Zahra, for the next generation of women in AI and tech leaders, what advice can you give?

Zahra: This is a very important question, one with many options, with many answers, but I will try to keep it short and objective for you. The AI and tech industry is a combination of art skills, but also there is an aspect of emotional intelligence. One must ensure that we have the perfect blend of both. This needs training. Once attained, believe me, it’s a superpower. 

I would advise women to have a five-year plan, know what your goals are, and work hard to achieve them. Take initiative, communicate, network to achieve them better, to achieve them faster. Go to conferences, go to meetups, get a mentor. Even with the current climate, this can still be done by virtual events. You will certainly connect with successful women in the industry to help support, to achieve success in your field. Don’t wait to be discovered and don’t blame others by saying they didn’t let me. Be proactive, not just reactive. 

Always remember—this is speaking from experience—you will fail. You will hear a thousand ‘no.’ But failure—to me at least—is not the opposite of success; it is part of success. I know I sound cliché but hearing another ‘no’ brings you one step closer to the magical ‘yes,’ where all doors of opportunity will be wide open. When that time comes—and definitely the time will come if you work hard enough—you will shine. 

When that time comes, be ready. Be yourself and remember that you worked hard for it, so you deserve it. Embrace that opportunity and enjoy it. Please remember, also help others when you get a chance to do so.

Jon: I do believe when you combine hard and soft skills to create a superpower, you are unstoppable. That’s great insight and a great point to end on. 

To recap today’s conversation with Zahra Timsah, MassMutual Life Insurance’s AI Governance and Ethics Lead, the next generation of AI-power products and services has the potential to transform every aspect of the health sector. From accelerated drug research through biophysical modeling, to improve diagnostic accuracy in medical imaging, to smart claims management with self-learning software. No corner of the $10 trillion global health industry will remain untouched. But these advances do not come without risks as biases in AI raise ethical concerns and present health risks to patients, as well as financial risk to providers and payers.

This episode has been part of our special series on women in AI. A big thanks Elizabeth Mitelman for organizing the series and joining the session today. That’s a wrap on today’s show. Thank you, Zahra, for joining us and FortressIQ for sponsoring. I’m Jon Knisley and this has been hello, Human.