Dying man creates AI voice to speak to wife after death 💬

Hey!

Welcome to this week’s edition of Supermedic, where we explore the latest developments in artificial intelligence and its transformative impact on healthcare.

Let’s get into it!

Victor

TODAY’S MENU

• Dying Man Creates AI Voice to Speak to His Wife After Death

• Apple Introduces 'Apple Intelligence' and New Health Features at WWDC 2024

• What Are AI Hallucinations and How to Fight Them?

• Google AI Transforms Wearable Data into Personalized Health Insights

• The AI Opportunity: How Payers Can Capture It Now?

Read time: under 6 minutes

AFTERLIFE

Dying Man Creates AI Voice to Speak to His Wife After Death

It might seem like something out of a Black Mirror episode, but it’s very real. Michael Bommer, a 61-year-old software designer from Berlin, diagnosed with terminal colon cancer, has found a unique way to leave a part of himself behind for his beloved wife, Anett.

With the help of Eternos.Life, Bommer is creating an interactive AI version of himself that Anett can turn to for comfort and conversation after he's gone.

How does it work?

• To build this digital memoir, Michael recorded 300 sentences to capture his voice and shared 150 stories spanning his life experiences, principles, and advice for future generations.

• The AI will draw from this collection to provide answers and engage in dialogue, much like Michael himself would.

While the couple acknowledges that this AI tool is not a path to immortality, they see it as a way for Bommer's wisdom and memories to live on. Anett envisions asking the AI to read poems, recount their first meeting, or relive the story of Bommer's proposal – a bittersweet way to remember their cherished moments together.

Ethical questions remain

Creating digital avatars of the deceased brings up issues of consent, privacy, and the psychological impact on the living. Some critics argue that relying on AI to maintain a connection with loved ones could hinder the natural grieving process, potentially causing more harm than good. Plus, the commercial exploitation of such technology raises concerns about commodifying personal memories and the authenticity of interactions with a digital version. As this technology evolves, it’s crucial to carefully navigate these ethical dilemmas to ensure it truly helps those it aims to support.

THEY SUPPORT US 🩵

Senior LatAm Tech Hiring in 24 Hours

CloudDevs is the largest pool of tech talent in LatAm. Our 10,000+ pre-vetted LatAm engineers have a minimum of 7 years of experience and are hand-selected for your project in just 24 hours.

Our talent is rigorously tested for problem-solving, tech stack and communication, and specialize in everything from AI and healthtech to fintech and web3.

Try CloudDevs free for 7 days – no risk, no obligation.

WEARABLE

Apple Introduces 'Apple Intelligence' and New Health Features at WWDC 2024

At the highly anticipated WWDC 2024, Apple introduced its new AI initiative, "Apple Intelligence." By partnering with OpenAI (ChatGPT), Apple aims to bring advanced AI capabilities to its devices, significantly enhancing the functionality of Siri. Alongside this, Apple also revealed a new app for Apple Watch called “Vitals”, making one more step into health monitoring.

Apple Intelligence brings Siri to life 🧠

Apple Intelligence marks a significant upgrade for Siri, enabling it to comprehend and interact with icons, widgets, and text on your screen. This allows users to ask Siri to summarize notifications, emails, meeting notes, and more, streamlining information management and task execution directly from the screen. It’s like having a seemless version of ChatGPT directly included in your native Apple apps (Mail, Photo, iMessage…). In beta for now, this major update will be available in early fall.

Apple Introduces the "Vitals" App for their Watches ⌚

The new watchOS 11 features the "Vitals" app, which monitors overnight health metrics, including heart rate, blood oxygen levels, and respiratory rate. It provides notifications for unusual readings and tracks health trends over time, promoting early awareness and proactive health management. This feature aligns closely with current health tracking technologies but leverages Apple's extensive user base and ecosystem (1 billion people on earth).

Implication for Healthcare

Apple's commitment in AI technologies opens new possibilities for healthcare applications. The integration of Apple Intelligence with health monitoring tools like the Vitals app could transform personal health management. Users could benefit from more accurate and timely health data, facilitating early detection of potential issues and more informed health decisions. As AI continues to evolve, the potential for enhanced patient care, remote monitoring, and personalized health interventions grows, positioning Apple as a key player in the digital health revolution.

Source: Apple Intelligence, WatchOS, WWDC highlights

LEARN ABOUT AI

What Are AI Hallucinations and How to Fight Them?

Have you ever noticed that sometimes AI agents like ChatGPT make things up out of nowhere? How and why does that happen? Is there a way to mitigate it? In this week's ‘Learn about AI’ section, we dive deep into these questions and explore practical solutions.

What are AI hallucinations?

AI hallucination is a phenomenon wherein a large language model (LLM) perceives patterns or objects that are nonexistent or imperceptible to human observers, creating outputs that are nonsensical or altogether inaccurate. In other words, an AI hallucination occurs when an AI model generates incorrect information but presents it as if it were a fact.

Why does it happen?

AI tools like ChatGPT have been trained to predict strings of words that best match your query. However, they lack the reasoning to apply logic or consider any factual inconsistencies in their responses. In short, AI hallucinations occur when the AI goes off the rails trying to please you.

How can we mitigate them?

These hallucinations can be critical in a medical environment, as they can lead to incorrect decisions and outcomes. As AI platforms like ChatGPT and Claude become more prevalent in healthcare, the risk of hallucinations remains a significant concern. Fortunately, there are ways to mitigate them.

Key strategies to follow:

• Ensure Human Oversight: Always have human input and oversight to verify AI-generated information. This ensures its accuracy and reliability.

• Use AI Observability Tools: Leverage tools that help detect anomalies, diagnose issues, and provide transparency into AI decision-making processes.

• Promote Education and Awareness: Stay informed about AI advancements (subscribe to Supermedic.ai for example 🙂). Encourage healthcare educational institutes to offer courses on "AI in healthcare" to prepare future professionals.

• Improve Data Accuracy: Remember, the quality of AI models depends on the accuracy of their training data. Focusing on data accuracy can significantly reduce hallucinations.

• Encourage Collaboration and Diversity: Foster ongoing collaboration between clinicians and AI experts, and use diverse datasets to enhance outcomes and eliminate biases.

AI in healthcare holds tremendous promise but must be used responsibly. By addressing the risks of hallucinations and ensuring human oversight, we can harness AI's full potential to improve patient care while safeguarding their well-being.

Source: IBM, Forbes

RESEARCH

Google's AI Agent Transforms Wearable Data into Personalized Health Insights

Wearable health devices collect a lot of data, but without proper interpretation, much of it goes unused. To change that, Google researchers have developed PHIA (Personal Health Insights Agent), an AI system designed to analyze data from wearable health trackers and provide personalized health insights.

Key Takeaways:

• Smart Data Analysis: PHIA can autonomously interpret data from fitness trackers and smartwatches, such as Fitbit, Apple Watch, and Garmin, offering clear and useful health insights.

• Advanced Reasoning: Combining code generation with web search, PHIA can answer complex health-related questions.

• High Accuracy: PHIA can correctly answer over 84% of factual numerical queries and 83% of open-ended questions, providing reliable and detailed health advice.

PHIA's ability to provide accurate, personalized health insights from wearable data has significant potential. This can help users develop personalized wellness plans, promoting healthier lifestyles. As this technology advances, it could revolutionize how we understand and manage personal health.

MANAGEMENT

The AI Opportunity: How Payers Can Capture It Now?

Healthcare payers are at a pivotal juncture, facing major economic pressures. However, there's a silver lining: AI and automation present a massive opportunity to lower costs and increase revenue.

A recent analysis from McKinsey found that by adopting the latest AI and generative AI technologies, payers could achieve:

• 13-25% savings in administrative costs

• 5-11% savings in medical costs

• 3-12% higher revenue

To fully capture this value, payers must fundamentally rewire their organizations across six key areas: business-led digital roadmap, talent, operating model, technology, data, and adoption and scaling.

CEOs need to ask tough questions and mobilize their entire organization to pursue this opportunity. While it can't happen overnight, the potential benefits are too significant to ignore.

Other AI Healthcare news :

• Radiology: AI-enhanced imaging software increases the accuracy of prostate cancer margins by 45 times.

• Cardiology: AI tool discovers 17 gene variants linked to heart disease, paving the way for personalized treatments.

• Neuroscience: Harvard and DeepMind create virtual rat model powered by AI neural network, to study brain control of movement.

• Gastroenterology: USC develops GPS-like smart pills with AI for real-time gut monitoring and early disease detection.

• Addictology: UC San Diego develops AI model to predict opioid addiction risk in patients

• Business: AI startup Regard secures $30M Series B funding to enhance clinical documentation and diagnosis support

Thanks for reading!

Let’s improve the newsletter together!🌟