AI Coach Helps Smokers Quit with Reinforcement Learning 🚬

Welcome to another edition of Horizon AI,

Researchers are constantly exploring new ways to apply AI in healthcare. This time, a researcher from the Netherlands has developed an AI-powered coach to help people quit smoking, a habit responsible for over 480,000 deaths each year.

Let’s jump right in!

Read Time: 4.5’ min

Here's what's new today in the Horizon AI

• Chart of the week: How US and China collaborate on AI despite rivalry

• Smart AI Coach Helps Smokers Quit Using Reinforcement Learning

• Free Resources

• AI tools to check out

• Video of the week

Chart of the week

How US and China collaborate on AI despite rivalry

• Data from the Emerging Technology Observatory at Georgetown University shows that the US and China are closely linked in AI research. In the past decade, their researchers have co-authored over 46,000 AI papers—more than any other country pair.

• By working together, researchers from both nations have advanced fields like machine learning and natural language processing.

• Many breakthroughs have come from collaborations between top US universities like Stanford and MIT and Chinese institutions like Tsinghua University.

Source

AI News

AI RESEARCH

Smart AI Coach Helps Smokers Quit Using Reinforcement Learning

Researcher Nele Albers at Delft University of Technology developed an AI coach that uses psychologically informed Reinforcement Learning (RL) to adapt to individual needs and offer personalized support to help smokers and vapers quit for good.

Details:

• RL is a type of machine learning where a model learns through rewards, similar to how humans learn behaviors.

• Albers based her research on insights from behavior change theories and data from three crowdsourcing studies with each more than 500 participants.

• The AI coach encourages users to reflect on the stimuli that trigger their desire to smoke or the person they want to become in the future to help them find motivation. Furthermore, it seeks a balance between smokers’ preferences, circumstances, and expert advice, making support more effective.

Albers concludes that her research shows “how personalizing support—by taking into account both a person's current and future state—increases the effectiveness of AI-based eHealth applications. This offers many opportunities for behavioral change."

Source

Resources

👀 What is Character AI? Everything to know about the role-playing AI tool

👉 Learn how to use and prompt OpenAI's o1 model for complex reasoning tasks – Short course

💡 Most US workers don't use AI at work yet. This study suggests a reason why

💊 Four ways to power-up AI for drug discovery

AI Tools to check out

🔍 Kraftful: Get insights to build products people love. AI summaries of user feedback from app store reviews, support tickets, call transcripts, and more.

📄 Speakdocs: Have natural conversations with your documents using AI-powered insights.

✈️ Flight Price Predictor: Know when to book flights without overpaying with reliable AI airfare predictions.

🎵 Eapy: It allows creators organize ideas, share inspiration and workflows, and co-create original, copyright-free music with 100% copyright-free AI music models.

⭐ Patterned AI: Generate unique patterns for your product using AI.

Video of the week

Generative AI's Greatest Flaw

Mike Pound from University of Nottingham explains in this video why indirect prompt injection is considered GenAIs greatest flaw,

He discusses how malicious instructions can be hidden in data sources used by LLMs, creating serious security risks such as data leaks or AI misuse. He also explores ways to reduce these risks, including carefully selecting data sources, thorough testing, and prompt analysis.

That’s a wrap!

Thanks for sticking with us to the end! Let’s stay connected on LinkedIn and Twitter.

Not subscribed yet? Sign up here and send it to a colleague or friend!

See you in our next edition!

Gina 👩🏻‍💻