How can AI detect cancer earlier?
In today's healthcare landscape, there is an urgent need for rapid and precise disease diagnosis. How can artificial intelligence revolutionise medical diagnostics and improve patient care?
In this TED x PwC film, we explore the journey of health tech innovator Neil Daly, a leader in the integration of AI within dermatology. By advancing AI technology, he aids specialist teams in identifying cancer more efficiently, thereby creating pathways to improved patient outcomes and transforming the future of healthcare.
I think this is a moment that we'll look back on as one of the most important and profound changes for human history.
How is AI transforming scientific research?Scott Likens: We’ve reached a point where AI is embedding insights into decision-making workflows, whether that’s through dashboards, real-time alerts, or predictive models. The vast amount of data collected over the decades can now be automated, curated, and fine-tuned by AI in ways that enrich our understanding and provide precise answers, allowing us to conduct and automate new research efficiently. It’s about building a culture where data and AI are part of the conversation, not just in the background. So, the possibilities are almost limitless.
How is AI influencing healthcare delivery, in terms of scientific discovery?Joe Atkinson: We're seeing AI detect patterns in organic data related to medical R&D, helping to untangle the big, complex, and interconnected areas like the nervous system, cardiology, and more. This technology allows us to not only advance new drug and treatment development rapidly but also streamline these processes, reducing research-to-market time and enabling businesses to be more connected to the market demands.
Looking to the future, how could AI change everyday, personal healthcare?Scott Likens: With AI and the data from our phones and wearables, we can have a continuous, honest view of our health, which opens up a remarkable period of personalised healthcare. AI-powered systems streamline everything from real-time interaction with doctors to mental health insights, making our healthcare intimately connected to the AI-driven approach.
What role is agentic AI going to play in this healthcare and research revolution? Bret Greenstein: Agentic AI brings intelligence to automation - providing the ability to reason and plan towards a solution. So agents can be trained on the actual process of research – how to form a hypothesis, collect data, validate or invalidate the hypothesis, change the hypothesis and go again. It’s just an extraordinary tool for scaled research, scaled experimentation and scaled simulation. They can actually ‘think’ like a researcher, so when you give them a goal, they will try and test that – but quickly, undistracted and less biased. Then when you connect those agents to learn from each other, suddenly we're able to do things at speed that humans couldn't do at all before – but always getting to an output the human is comfortable with.
Where are we already seeing agentic AI being used today?Bret Greenstein: We're seeing it, obviously, in drug development. Agents are being used to explore different chemical and material compositions, and their effects. This has always been a huge research challenge – understanding reactions on the human body. Can you evaluate enough of those variables to see what’s helpful without being harmful? Now with the ability to manage bigger data sets and detect patterns – without the physical testing – there's no question we're going to accelerate drug discovery and development in many different places. Another exciting place is in radiology – evaluating what we're seeing in the scans, MRIs and all the other places where we’re taking data. That’s traditionally been a very manual, expensive, highly specialised activity. But if you get the agentic AI to do more of the data gathering and summarising, then you can apply human expertise to the most difficult parts of the problem – improving insights and outcomes.
What does this immense power mean for human researchers and healthcare professionals?Joe Atkinson: AI will always be there to support the human, not supplement the human. For example, AI on its own isn’t going to be able cure cancer or solve world hunger, but the people who have these tools now have more capability than at any time in history. And today, healthcare organisations are wrapped up in complex, time-consuming processes – from research to compliance, administration, scheduling and facilities management. All these are areas where AI can actually improve the patient experience and healthcare delivery, while also saving money. And you’re not even replacing talent, because there’s already a shortage of talent in these areas. So organisational design will fundamentally change as a result of AI agents, but that's not a bad thing – organisations have been redesigned over the decades.
What would your advice be to leaders who are leading an innovation team? Joe Atkinson: Organisations that truly reinvent their approaches with AI are poised to achieve much better results and much higher returns. The key lies in rethinking, reinventing, and reimagining the end-to-end process, rather than merely aiming for greater efficiency. As AI evolves rapidly, it's crucial for organisations to balance this pace of change with their core mission and ethics. This involves clearly understanding the rapid technological advancements and upskilling the workforce to pre-emptively adapt to these changes. In the short term, the focus may be on cost savings and keeping pace, but long-term success requires investment in people and technology. Integrating AI into both operations and innovation fosters shared victories and cultivates a culture that celebrates experimentation while clearly defining the human role and differentiator amidst evolving AI agents.
Organisations that truly reinvent their approaches with AI are poised to achieve much better results and much higher returns.
How can agentic AI be incorporated in a responsible way?Brenda Vethanayagam: When we think about AI agents, people often refer to them as an extension of the workforce – like extra employees. Just as we train people to meet objectives responsibly, AI agents require similar training with appropriate guardrails and validation. This means that they operate according to the values and parameters with which you're comfortable. Practically, this involves engaging everyone in the organisation who’s part of the AI journey, ensuring alignment on strategy, objectives, and training. For instance, if the strategy prioritises rapid adoption, processes, risk management, technology, and training must adapt accordingly. AI agents need embedded validations and aligned values, and involving all organisational stakeholders leading to a cohesive AI journey. Responsible AI incorporates transparency, governance, and processes, fostering a responsible culture where AI operability is fully understood.
How do we build and maintain trust in AI, while unlocking the upsides and accelerating breakthroughs?Brenda Vethanayagam: Firstly, it is crucial to place human safety at the centre of all innovation efforts. Trust, a profoundly human trait, requires time to develop and is nurtured through honesty, transparency, demonstrating AI capabilities, and delivering tangible results. Much of the fear surrounding AI stems from uncertainty, particularly concerning loss of control, privacy, and data security. To mitigate these concerns, it is important to clarify the ‘design principles’—how implementations occur, the guardrails and controls in place, and precisely what is being done. Clear communication about these elements helps ease fears associated with privacy and control. Over time, the outcomes, such as breakthroughs in AI leading to better treatments and saved lives, will foster trust in its capabilities. With human safety consistently at the forefront, these developments will rapidly alleviate fears, encouraging inclusion and generating lasting impacts.
