He says that: AI is especially important for biopharma, because AI can understand biology in a way that humans cannot; and Biopharma is in a particularly strong position to adopt AI, because it is simultaneously highly profitable and extremely inefficient (given the iterative trial and error approach towards drug discovery). Here’s a bit more on each of those points. 1. AI is especially important for biopharma The reason AI is likely to have such a profound impact in biopharma is that the same sort of transformer technology that can (sort of) understand natural languages like English can also be applied to long molecules like proteins and RNA. Humans understand and can manipulate English better than AI can; by contrast the way molecules work is not intuitive to humans. As a result, it’s fair to say that in many ways AI understands the language of large bio-molecules better than humans do. One sign of AI’s power to change biopharma is that several of the individuals who have been at the cutting edge of AI are now focusing on using AI to accelerate the development of drugs – for example, Demis Hassabis (the founder and CEO of DeepMind) and Daphne Koller (a former professor of computer science at Stanford). Another indication is that of all the academic papers on AI, the ones that have been most cited in the past couple of years have actually focused on proteins. 2. Biopharma is also in a very strong position to adopt AI Biopharma operates a business model quite unlike any other industry, and many of the peculiarities seem tailor-made to adopting AI: Lots of cash to invest in AI. When drugs are approved they operate as quasi- monopolies, with very high margins. This means both that many pharma companies have strong cash flows, and that there are huge levels of R&D investment in the biopharma sector. There is plenty of scope for AI to bring improvements because the industry involves extraordinary inefficiencies. Developing new drugs often takes 15 years and costs billions of dollars. Typically, thousands of molecules are tested before one is taken to clinical trials – but even then 92% of the candidate drugs fail in clinical trials. For those drugs that succeed, it takes an average of more than a decade from the start of clinical trials to approval. And time is of the essence. Biotech companies have to file patents on their drugs before they start clinical trials, and the patents last for only 20 years, at least in theory. Even though they are often extended in practice, pharma companies have only a limited number of years to generate returns from the drugs. Anything that can bring drugs to market faster is therefore likely to have a major economic impact – and there is plenty of evidence that AI should be able to do that. AI is likely to improve the industry’s cashflow: New drugs can generate revenues in the billions. If AI allows new drugs to be found – to drug the previously undruggable – then the upside could potentially be worth billions of dollars. Bringing drugs to market faster can improve cash flows very substantially. AI can produce drugs more quickly, which means the drug is likely to produce revenue for a greater percentage of the patent period. Accelerating approval by 12 months can increase the NPV of a drug by $100 million. Better drugs can generate higher revenues. AI-designed drugs have the potential for greater revenue, even if they are treating diseases where therapies already exist, because they can be targeted more tightly and designed with fewer side effects. AI can reduce the cost of development substantially, especially in the discovery phase. Exscientia says AI-driven drug development can reduce the costs in the preclinical phases by about 70%, relative to traditional techniques. Biopharma is the largest subsector within healthcare in economic terms One reason why the impact of AI on biopharma is important is that it is the dominant subsector of healthcare in financial terms. Biopharma accounts for very roughly half of the industry’s profit and investment, even though prescription drugs account for only 11% of spending on healthcare in the US (see first chart below). The second chart below shows that over the last two years, $102 billion was invested in biotech and drug start-ups, and a further $7 billion was invested in tools for drug discovery, together making up 48% of total investment by VCs in healthcare. US Healthcare Spend by Category, 2021 Source: CMS Primary Capital Raised in Healthcare Startups, 2021 and 2022 combined ($ in Blns) Source: PitchBook The full report goes on to look at companies that are using AI in biopharma. Colleagues in Citi Global Data Insights (CGDI) are able to respond to requests to generate screens of companies exposed to the themes in this report, based on a quantitative analysis both of patents filed and of newsflow. These screens can cover quoted and unquoted companies. They can be ranked in many ways. For example, it is possible to rank by the number of patents obtained related to AI in biopharma, or the quality of those patents, or the percentage of a company’s patents that fall in this area. Please reach out to the author (adam.spielman@citi.com) or to Helen Krause (helen.krause@citi.com) at CGDI for more information on any of this. Please find the full report here, originally published on 23 June 2023, and if you are a Velocity subscriber: https://www.citivelocity.com/ And please find the previous reports in the series here: Smart Thinking on AI in Healthcare: Part 1 – Overview Smart Thinking on AI in Healthcare: Part 2 – Medical Devices Smart Thinking on AI in Healthcare: Part 3 – Health Administration Citi Global Insights (CGI) is Citi’s premier non-independent thought leadership curation. It is not investment research; however, it may contain thematic content previously expressed in an Independent Research report. For the full CGI disclosure, click here.

He says that:
<ul>
<li>AI is especially important for biopharma, because AI can understand biology in a way that humans cannot; and</li>
</ul>
<ul>
<li>Biopharma is in a particularly strong position to adopt AI, because it is simultaneously highly profitable and extremely inefficient (given the iterative trial and error approach towards drug discovery).</li>
</ul>
Here&rsquo;s a bit more on each of those points.
1. AI is especially important for biopharma
The reason AI is likely to have such a profound impact in biopharma is that the same sort of transformer technology that can (sort of) understand natural languages like English can also be applied to long molecules like proteins and RNA. Humans understand and can manipulate English better than AI can; by contrast the way molecules work is not intuitive to humans. As a result, it&rsquo;s fair to say that in many ways AI understands the language of large bio-molecules better than humans do.
One sign of AI&rsquo;s power to change biopharma is that several of the individuals who have been at the cutting edge of AI are now focusing on using AI to accelerate the development of drugs &ndash; for example, Demis Hassabis (the founder and CEO of DeepMind) and Daphne Koller (a former professor of computer science at Stanford).
Another indication is that of all the academic papers on AI, the ones that have been most cited in the past couple of years have actually focused on proteins. &nbsp;&nbsp;
2. Biopharma is also in a very strong position to adopt AI
Biopharma operates a business model quite unlike any other industry, and many of the peculiarities seem tailor-made to adopting AI:
<ul>
<li>Lots of cash to invest in AI. When drugs are approved they operate as quasi- monopolies, with very high margins. This means both that many pharma companies have strong cash flows, and that there are huge levels of R&amp;D investment in the biopharma sector.</li>
</ul>
<ul>
<li>There is plenty of scope for AI to bring improvements because the industry involves extraordinary inefficiencies. Developing new drugs often takes 15 years and costs billions of dollars. Typically, thousands of molecules are tested before one is taken to clinical trials &ndash; but even then 92% of the candidate drugs fail in clinical trials. For those drugs that succeed, it takes an average of more than a decade from the start of clinical trials to approval. </li>
</ul>
<ul>
<li>And time is of the essence. Biotech companies have to file patents on their drugs before they start clinical trials, and the patents last for only 20 years, at least in theory. Even though they are often extended in practice, pharma companies have only a limited number of years to generate returns from the drugs. Anything that can bring drugs to market faster is therefore likely to have a major economic impact &ndash; and there is plenty of evidence that AI should be able to do that.</li>
</ul>
AI is likely to improve the industry&rsquo;s cashflow:
<ul>
<li>New drugs can generate revenues in the billions. If AI allows new drugs to be found &ndash; to drug the previously undruggable &ndash; then the upside could potentially be worth billions of dollars.</li>
</ul>
<ul>
<li>Bringing drugs to market faster can improve cash flows very substantially. AI can produce drugs more quickly, which means the drug is likely to produce revenue for a greater percentage of the patent period. Accelerating approval by 12 months can increase the NPV of a drug by $100 million.</li>
</ul>
<ul>
<li>Better drugs can generate higher revenues. AI-designed drugs have the potential for greater revenue, even if they are treating diseases where therapies already exist, because they can be targeted more tightly and designed with fewer side effects.</li>
</ul>
AI can reduce the cost of development substantially, especially in the discovery phase. Exscientia says AI-driven drug development can reduce the costs in the preclinical phases by about 70%, relative to traditional techniques.
<h1>Biopharma is the largest subsector within healthcare in economic terms</h1>
One reason why the impact of AI on biopharma is important is that it is the dominant subsector of healthcare in financial terms. Biopharma accounts for very roughly half of the industry&rsquo;s profit and investment, even though prescription drugs account for only 11% of spending on healthcare in the US (see first chart below).
The second chart below shows that over the last two years, $102 billion was invested in biotech and drug start-ups, and a further $7 billion was invested in tools for drug discovery, together making up 48% of total investment by VCs in healthcare.
US Healthcare Spend by Category, 2021
<img src="/rcs/v1/media/citigpa/akpublic/storage/retrieveMediaFile/64ef6343860feb324c9fed5c.png" width="630" height="434" />
Source: CMS 
Primary Capital Raised in Healthcare Startups, 2021 and 2022 combined ($ in Blns) 
<img src="/rcs/v1/media/citigpa/akpublic/storage/retrieveMediaFile/64ef63a4daa69a4e782489ef.png" width="631" height="420" />
Source: PitchBook
The full report goes on to look at companies that are using AI in biopharma. Colleagues in Citi Global Data Insights (CGDI) are able to respond to requests to generate screens of companies exposed to the themes in this report, based on a quantitative analysis both of patents filed and of newsflow.
These screens can cover quoted and unquoted companies. They can be ranked in many ways. For example, it is possible to rank by the number of patents obtained related to AI in biopharma, or the quality of those patents, or the percentage of a company&rsquo;s patents that fall in this area.
Please reach out to the author (<a href="mailto:adam.spielman@citi.com" target="_blank" rel="noopener" data-link-source="tiny">adam.spielman@citi.com</a>) or to Helen Krause (<a href="mailto:helen.krause@citi.com" target="_blank" rel="noopener" data-link-source="tiny">helen.krause@citi.com</a>) at CGDI for more information on any of this.
Please find the full report here, originally published on 23 June 2023, and if you are a Velocity subscriber: <a href="https://www.citivelocity.com/" target="_blank" rel="noopener" data-link-source="tiny">https://www.citivelocity.com/</a>
And please find the previous reports in the series here:
<a href="https://www.citivelocity.com/rendition/eppublic/public/documentService/ZG9jX2lkPTMwMTQzMjE5" target="_blank" rel="noopener" data-link-source="tiny">Smart Thinking on AI in Healthcare: Part 1 &ndash; Overview</a> 
<a href="https://www.citivelocity.com/rendition/eppublic/public/documentService/ZG9jX2lkPTMwMTQ1MjEz" target="_blank" rel="noopener" data-link-source="tiny">Smart Thinking on AI in Healthcare: Part 2 &ndash; Medical Devices</a> 
<a href="https://www.citivelocity.com/rendition/eppublic/public/documentService/ZG9jX2lkPTMwMTQ4MjY0" target="_blank" rel="noopener" data-link-source="tiny">Smart Thinking on AI in Healthcare: Part 3 &ndash; Health Administration</a> 
Citi Global Insights (CGI) is Citi&rsquo;s premier non-independent thought leadership curation. It is not investment research; however, it may contain thematic content previously expressed in an Independent Research report. For the full CGI disclosure, click <a href="https://ir.citi.com/6A3IURUaezPxFLppsgdBkqidjLaLn558JwHISYaDZcTfE9XMRB1isDDDVorqrxpR%2fTnEi4RUCl8%3d" target="_blank" rel="noopener" data-link-source="tiny">here</a>.

A recent report from Adam Spielman of Citi Global Insights looks at biopharma and why AI is especially important in this area. He says biopharma will be affected by AI more than any other aspect of the healthcare industry. In fact, he says it will be affected more than almost any other industry. 

AI in Healthcare: The Importance of Biopharma 

AI tools to automate diagnosis and prescription. The report says AI will initially be used by doctors to ensure they haven’t missed anything. Over time, we expect AI will start analyzing data in ways that no human doctor could. AI agents to provide services currently available only from healthcare professionals. AI agents can now pass the professional exams required for doctors, nurses, dietitians, and similar healthcare professionals, but the cost of providing an AI agent is around 20¢ an hour, vs. $20 to $90 an hour for a human nurse, dietitian, or healthcare coder in the US. AI health agents can therefore provide services in relatively low-risk situations that are uneconomic to provide with human professionals. An AI agent could be used, for example, to generate weekly (or daily) calls to check in on elderly individuals, gently reminding them to take their medicines; reassuring them about minor health issues; and, if needed, booking appointments with (human) healthcare professionals and arranging the necessary transport. AI for consumers that replicates the conversation of doctors: Already most people perform an Internet search of health issues before they visit a clinician. In the future, their health questions will be answered by large language models (LLMs), and these will provide answers that sound very much like (friendly and well-informed) doctors. These changes will take time, the report says, but on a 5- to 10-year view these tools -- and others like them -- will change the role of most healthcare professionals, altering the way healthcare is delivered very substantially. Tools for automating diagnosis Increasingly effective tools for diagnosis are expected to emerge, formed from the combination of four increasingly mature technologies, each of which is powered by generative AI: Speech-to-text: Automatic speech recognition (ASR) now allows smart speakers to transcribe accurately what a patient is saying, in the clinic or during a phone call. The technology has improved significantly since generative AI has become available, because even when the software cannot distinguish the word, AI makes well-educated guesses about what the speaker was trying to say. Generative AI can extract the main points from the transcript: LLMs are less likely to hallucinate when they’re asked to summarize text than when they’re asked to create something entirely new. Medically specialized LLMs can suggest next steps as effectively as human doctors. Some LLMs – for example Google’s Med-PaLM2 – can outperform human doctors in medical exams. These effectively ask the examinee – whether human or machine – to analyze medical vignettes. But we’ve just said in bullet points 1 and 2 above that AI can reduce conversations with real humans to their essentials. Text-to-speech: Finally, systems now exist to convert the text created in step 3 back to speech. The report says that, in the first place, diagnosis tools will be used by doctors to check they haven’t missed something. For example, a family doctor, listening to a patient, may think it’s worth investigating A, B, and C; however, the AI may also remind the doctor that syndromes D and E are also possible and therefore need consideration. As a result, these tools should help doctors provide a more consistent service, speeding up their work, and also remove some of the stress from the job. In China, AI is already used by more than a more than 1 million doctors to help diagnose faster and more accurately. The report says its popularity reflects in part the underdeveloped state of China’s system of family doctors, who often have many years less training than their western equivalents. The system is particularly popular in smaller cities, where often there is only a rudimentary primary care system. This is the fifth in a series on AI in healthcare; details in the table below. Reports in the Smart Thinking on AI in Healthcare series © 2023 Citigroup Inc. No redistribution without Citigroup’s written permission. Source: Citi Global Insights The full report features greater depth on specific AI tools for healthcare professionals and how AI diagnoses are likely to start outperforming human doctors over time and much more. For more information on this subject, and if you are a Velocity subscriber, please see the full report, originally published on 18 August 2023, here: https://www.citivelocity.com/ Citi Global Insights (CGI) is Citi’s premier non-independent thought leadership curation. It is not investment research; however, it may contain thematic content previously expressed in an Independent Research report. For the full CGI disclosure, click here.

AI tools to automate diagnosis and prescription. The report says AI will initially be used by doctors to ensure they haven&rsquo;t missed anything. Over time, we expect AI will start analyzing data in ways that no human doctor could.
AI agents to provide services currently available only from healthcare professionals. AI agents can now pass the professional exams required for doctors, nurses, dietitians, and similar healthcare professionals, but the cost of providing an AI agent is around 20&cent; an hour, vs. $20 to $90 an hour for a human nurse, dietitian, or healthcare coder in the US. AI health agents can therefore provide services in relatively low-risk situations that are uneconomic to provide with human professionals. An AI agent could be used, for example, to generate weekly (or daily) calls to check in on elderly individuals, gently reminding them to take their medicines; reassuring them about minor health issues; and, if needed, booking appointments with (human) healthcare professionals and arranging the necessary transport.
AI for consumers that replicates the conversation of doctors: Already most people perform an Internet search of health issues before they visit a clinician. In the future, their health questions will be answered by large language models (LLMs), and these will provide answers that sound very much like (friendly and well-informed) doctors.&nbsp;
These changes will take time, the report says, but on a 5- to 10-year view these tools -- and others like them -- will change the role of most healthcare professionals, altering the way healthcare is delivered very substantially.
<h1>Tools for automating diagnosis</h1>
Increasingly effective tools for diagnosis are expected to emerge, formed from the combination of four increasingly mature technologies, each of which is powered by generative AI:
<ul>
<li>Speech-to-text: Automatic speech recognition (ASR) now allows smart speakers to transcribe accurately what a patient is saying, in the clinic or during a phone call. The technology has improved significantly since generative AI has become available, because even when the software cannot distinguish the word, AI makes well-educated guesses about what the speaker was trying to say.</li>
</ul>
<ul>
<li>Generative AI can extract the main points from the transcript: LLMs are less likely to hallucinate when they&rsquo;re asked to summarize text than when they&rsquo;re asked to create something entirely new. </li>
</ul>
<ul>
<li>Medically specialized LLMs can suggest next steps as effectively as human doctors. Some LLMs &ndash; for example Google&rsquo;s Med-PaLM2 &ndash; can outperform human doctors in medical exams. These effectively ask the examinee &ndash; whether human or machine &ndash; to analyze medical vignettes. But we&rsquo;ve just said in bullet points 1 and 2 above that AI can reduce conversations with real humans to their essentials.&nbsp;</li>
</ul>
<ul>
<li>Text-to-speech: Finally, systems now exist to convert the text created in step 3 back to speech.</li>
</ul>
The report says that, in the first place, diagnosis tools will be used by doctors to check they haven&rsquo;t missed something. For example, a family doctor, listening to a patient, may think it&rsquo;s worth investigating A, B, and C; however, the AI may also remind the doctor that syndromes D and E are also possible and therefore need consideration.
As a result, these tools should help doctors provide a more consistent service, speeding up their work, and also remove some of the stress from the job.
In China, AI is already used by more than a more than 1 million doctors to help diagnose faster and more accurately. The report says its popularity reflects in part the underdeveloped state of China&rsquo;s system of family doctors, who often have many years less training than their western equivalents. The system is particularly popular in smaller cities, where often there is only a rudimentary primary care system.
This is the fifth in a series on AI in healthcare; details in the table below.
Reports in the Smart Thinking on AI in Healthcare series
<img src="/rcs/v1/media/citigpa/akpublic/storage/retrieveMediaFile/64ef6890860feb324c9fed6d.png" />
&copy; 2023 Citigroup Inc. No redistribution without Citigroup&rsquo;s written permission. 
Source: Citi Global Insights
The full report features greater depth on specific AI tools for healthcare professionals and how AI diagnoses are likely to start outperforming human doctors over time and much more.
For more information on this subject, and if you are a Velocity subscriber, please see the full report, originally published on 18 August 2023, here: <a href="https://www.citivelocity.com/" target="_blank" rel="noopener" data-link-source="tiny">https://www.citivelocity.com/</a>
Citi Global Insights (CGI) is Citi&rsquo;s premier non-independent thought leadership curation. It is not investment research; however, it may contain thematic content previously expressed in an Independent Research report. For the full CGI disclosure, click <a href="https://ir.citi.com/6A3IURUaezPxFLppsgdBkqidjLaLn558JwHISYaDZcTfE9XMRB1isDDDVorqrxpR%2fTnEi4RUCl8%3d" target="_blank" rel="noopener" data-link-source="tiny">here</a>.
&nbsp;
&nbsp;
&nbsp;

A recent report from Adam Spielman of Citi Global Insights focuses on how the role of doctors will change with AI. Generative AI tools are set to increasingly automate many aspects of healthcare. What will the impact be on doctors and other clinicians? 

AI in Healthcare: The Role of Clinicians  

Insights: AI in Healthcare: Health Administration A recent report from Adam Spielman of Citi Global Insights looks at the possibilities for AI to impact health administration, potentially cutting costs and helping ease the huge burden on clinicians. Insights: AI in Healthcare: The Importance of Biopharma A recent report from Adam Spielman of Citi Global Insights looks at biopharma and why AI is especially important in this area. He says biopharma will be affected by AI more than any other aspect. Insights: AI in Healthcare: The Role of Clinicians A recent report from Adam Spielman of Citi Global Insights focuses on how the role of doctors will change with AI. Generative AI tools are set to increasingly automate many aspects of healthcare.

<h1><a href="https://www.citigroup.com/global/insights/global-insights/ai-in-healthcare-health-administration-" target="_blank" rel="noopener" data-link-source="tiny">Insights: AI in Healthcare: Health Administration</a></h1>
A recent report from Adam Spielman of Citi Global Insights looks at the possibilities for AI to impact health administration, potentially cutting costs and helping ease the huge burden on clinicians.
<h1><a href="https://www.citigroup.com/global/insights/global-insights/ai-in-healthcare-the-importance-of-biopharma-" target="_blank" rel="noopener" data-link-source="tiny">Insights: AI in Healthcare: The Importance of Biopharma</a></h1>
A recent report from Adam Spielman of Citi Global Insights looks at biopharma and why AI is especially important in this area. He says biopharma will be affected by AI more than any other aspect.
<h1><a href="https://www.citigroup.com/global/insights/global-insights/ai-in-healthcare-the-role-of-clinicians-" target="_blank" rel="noopener" data-link-source="tiny">Insights: AI in Healthcare: The Role of Clinicians</a></h1>
A recent report from Adam Spielman of Citi Global Insights focuses on how the role of doctors will change with AI. Generative AI tools are set to increasingly automate many aspects of healthcare.

Citi’s Adam Spielman is the author of a series of reports on the impact of AI on the healthcare sector.  Here he discusses several aspects of his research including AI’s impact on clinicians, its potential to drastically lessen their administrative burden, and why biopharma could be the area that AI impacts most of all. 

AI in Healthcare: Are AI Doctors on Call? 

Anne Wojcicki, CEO of 23andMe, talks to Citi&rsquo;s Adam Spielman about how genetic data could hold the key to a more preventative, accessible and personal approach to healthcare.