The Dr. Google does exist, yes, but it is not a hasty search before going to the doctor. It is a woman, like 70% of healthcare professionals. Her name is Susan Thomas, she is English, and she serves as the director of Google Health, the healthcare branch of the tech giant. After studying Medicine, she specialized in Geriatrics and practiced for a few years in Leeds before venturing into the world of consulting: after obtaining an MBA in Sydney and Chicago, she spent 15 years at Ernst & Young, an experience that she has now poured into the multinational's projects. Throughout the conversation with Thomas, who received us during her recent visit to Spain (her first visit to our country), the pragmatism that she perceives in all scientific-health improvements is evident: "Technology helps us, but it does not replace us."
Question. Google constantly faces the challenge of new technologies and how they impact the present and future of people's health. What are the current challenges?
Answer. It is best to start by reflecting on what we do in Google Health. This part of the company represents our effort to promote millions of healthier lives. This statement may sound very ambitious, but in reality, when you think about what Google is, it makes sense. Every day, hundreds of millions of people turn to our services for health information. The Doctor Google is there.
Q. In some cases, the term Dr. Google has a negative connotation. From the area you lead, how do you give it a more constructive meaning?
A. People use both Google search and YouTube through their devices to obtain health information. It makes sense that, as an information company, our mission is to organize the world's information to make it more accessible and useful for everyone, including in health. Right now, we are very focused on three areas where artificial intelligence (AI) is the common element: how we make it more useful for consumers, how we work with healthcare professionals, and how we make it a support element for scientific research.
Q. Let's start with how it affects citizens who use the search engine to inform themselves about their health.
A. For many years, we have had what we call knowledge panels in the search engine. These are the initial sections that appear if you search, for example, diabetes, asthma, or depression. It is a box that contains authorized and verified information.
Q. How do you ensure it meets quality standards?
A. It has links to authorized sources. We have partnered with the NHS, the US National Academy of Medicine, the Mayo Clinic, and the WHO so that information can be trusted, and one can know they are consulting the right place. However, we have only been able to cover a certain number of diseases and in a certain number of languages. In the last two years, we have been developing what we call our AI summaries, where our Gemini models extract information from top searches and synthesize it. This way, information on many more conditions can be displayed and other questions answered.
Q. Does Gemini (your AI) already respond in natural language? Is it no longer necessary to search for concepts?
A. The way people ask questions has changed. Instead of just typing diabetes, someone might ask, "How do I manage my diabetes with glucose monitors?" Gemini will be able to extract specific information and provide it. We are working very hard to ensure that the information is as useful as possible. Our models are grounded to avoid inventing things.
Q. So, have you addressed the issues of health information supply? Do you verify that there is no misinformation or errors on all your platforms?
A. We are really confident that people can now trust Doctor Google with both knowledge panels and AI summaries. Both are available in Spain. We also have the same on YouTube. If you search for health information on this platform, authorized health sources such as doctors, nurses, hospitals, healthcare providers, appear at the top.
Q. And what do you do with misinformation?
A. Misinformation is actively removed. For example, we do not want people to find things written by someone sitting in their basement inventing theories about diabetes.
Q. How do you control and prioritize health information on YouTube?
A. We consider that many people, especially young ones, turn to this platform. Here we have formed communities of medical creators, nurses, and psychologists who produce videos on common health conditions that are really important. Whether it's acne for teenagers and how to deal with it or female health issues. There are certain general practitioners who already specialize in this because people feel a bit embarrassed to ask in the consultation: it's much easier to watch it in a video. So, again, we try to build a corpus of very high-quality information through our services.
Q. You mentioned at the beginning that this is not just consumer information, but you also seek to ally with Medicine providers, healthcare professionals. What do you offer them?
A. We work on creating products that make their daily work easier to provide quality care. For example, if you think about going to see your primary care doctor, you may only have 10 minutes with them because we know they see more patients than their capacity allows. These doctors have patient histories, records of each appointment, either on paper (in some places they still are) or on the screen. The issue is that you arrive, sit down, and wait while they spend the first half of the appointment looking at the history. Now we can use Gemini, our language model, to summarize the information and present it to the doctor before they sit in front of the patient.
Q. How does this assistant work?
A. Instantly, in a matter of seconds, it will give the doctor an accurate summary obtained from the history with just one click. At all times, you can check where each note comes from to verify its accuracy, as it is based on the record or history. This means that if they are seeing a patient, instead of spending only half the time listening to them and talking to them about what is happening, they can dedicate the full 10 minutes. This is an advantage for the doctor because they do not have to worry and have the cognitive burden of trying to cover everything. And it is much better for the patient because they do not feel that time is scarce. We also work with the same system for nursing.
Q. What projects or where have you implemented the system?
A. We have partnered with the HCA hospital group in the US. Here, they use Gemini for nursing handovers. Instead of having to write everything, they can extract nursing notes from medical records as concise summaries. This leads us to focus on the quality of information and patient safety.
Q. How would it work in Nursing?
A. If you are a nurse in charge, for example, of a ward with 25 patients, there is a lot to remember. And if you were just called and didn't write it down, you have lost information. While Gemini helps them deal with that. There is much work to see how we turn the need into a solution.
Q. The AI tool, being an information element, becomes a help for doctors' bureaucratic workload?
A. We aim to reduce administrative burden. There are many forms that doctors and nurses have to fill out, and it takes them a lot of time. Errors often slip in. So, in patient summaries during their admission, for example, there is information on why they came, how they were treated, the medication they are going home with...
Q. Let's now move to the scientist's side: what does Google Health offer to a laboratory researcher?
A. There are so many things... At the beginning of the month, we published an article about our AI co-scientist, which is a kind of set of agents driven by Gemini to help research scientists unlock novel hypotheses.
Q. Have you achieved good results?
A. In one of the projects, we worked with Imperial College. From this institution, Dr. Tiago Dias da Costa told us that he thought his hard drive had been hacked because our co-scientist came up with ideas in two days that took the researcher three years to develop and had kept hidden from us. This shows how findings are boosted; it's not about replacing them but simply accelerating what they are capable of doing.
Q. Are there any projects left unmentioned?
A. Yes, a very early one. It is AMIE (Articulated Medical Intelligence Explorer), which is a chatbot that patients can ask questions to. It is clear that this is the path of the future: in five or 10 years, we will all have a doctor in our pocket whose way of interacting with us will change clinical care. We will have personal health agents to guide us.
Q. However, we must not forget the digital divide because there are many generations that are not digital natives. How do you address this issue?
It is true: we must ensure that when we think about products, we consider different generations. Older people are different and when it comes to health, they need to feel the human touch in their interactions. We must also ensure that they do not become technologically isolated, either due to their age or economically unfavorable situations.
Q. All that you manage revolves around especially sensitive data and information, often of patients. How do you ensure that they are not misused?
The answer is simple: your data is yours, and it is only shared where you choose. Google does not want your data. It may be used in the models we have developed to provide you with more information. But it does not feed the models, not unless you consent. The same applies at a medical level.