Mental Health - GPT
Can you solve it with AI?
An inbox designed for reading newsletters: Meco is a distraction-free space for reading newsletters outside the inbox. Become a more productive reader and focus on what matters to you - easy to try with a free version. Try today
Friends,
I’m just off the back of London Edtech Week, which gives an up close and personal view of the pace of change in the industry. And while, parents know that AI is on the agenda in schools in different ways, 50% of teachers have used AI in lesson prep. The next-gen are going to be brought up with AI, much like the current gen have been brought up with socials.
However, other industries are seeing faster larger valuation-type impact from generative AI and I’m always interested in what’s going on elsewhere. Tech businesses and consumer electronics see a massive direct impact from what AI brings, but even healthcare and biotech see shifts, from all angles.
Companies like Viz.ai, Eko, Prognos, Freenome and Butterfly are all changing the face of what is possible with the way that data is used. I’m no expert, but cancer prevention, early diagnosis, early monitoring are all the type of technology applications that will increase healthspan and lifespan. These and other companies like them are at the forefront of integrating AI into healthcare, driving innovation in medical diagnostics, treatment planning, and personalized medicine. Their contributions are transforming the way healthcare is delivered, making it more accurate, efficient, and accessible. We don’t see these advancements. Inevitably, they start with tests, then in private care, and then government adopt them.
AI applications will have growing financial impact in medical imaging, diagnostics, and biotechnology research. And we forget about AI’s role in helping to solve the largest crisis in modern times: Moderna utilised AI to accelerate the development of its COVID-19 vaccine, mRNA-1273 significantly. Leveraging its existing AI and machine learning platforms, Moderna rapidly identified the genetic sequence of the SARS-CoV-2 virus and designed the corresponding mRNA sequence for the vaccine. AI algorithms were employed to optimize the mRNA structure for stability and efficacy, ensuring the vaccine would effectively prompt an immune response. The use of AI also facilitated the rapid synthesis and testing of mRNA constructs, allowing Moderna to move from sequence selection to clinical trials in record time. At the time it was a race to save the world - and I didnt really care about the how. But Moderna flipped the switch in how it operated. By integrating AI into their development process, Moderna was able to streamline the typically lengthy vaccine development timeline, demonstrating the transformative potential of AI in biotechnology and medical research. This innovative approach not only expedited the availability of a crucial vaccine but also set a new standard for future vaccine development processes.
Imagine if we could do this elsewhere. Imagine if we could solve mental health, another major issue of our time.
And this week I heard about Woebot (US Only), a product to try to do just that. Woebot is an AI-driven chatbot designed to provide mental health support through cognitive-behavioural therapy (CBT) techniques. It engages users in therapeutic conversations, offering personalized coping strategies and mood tracking to help manage anxiety, depression, and other mental health issues. Woebot's uniqueness lies in its accessibility, delivering immediate, evidence-based mental health support anytime, and its ability to create a supportive and interactive experience that mimics human empathy through advanced natural language processing.
It’s not a new product - but it brings the ethos of solving problems using AI. Like we would solve disease, or cure for cancer, or solving for improving children’s grades. Get the data, crunch the numbers, refine, change the model, and retest.
Whether it’s woebot or something else - I wonder if by thinking simply about problems, we should try and break them down to their basic level. For example, all Woebot did was check in daily with users, track their progress, and offer continuous support - but then jumped on Generative AI when given the opportunity.
In exploring the potential of LLMs, Woebot's team found generative AI could enhance emotional engagement but often fell short in delivering structured, therapeutic interactions essential for CBT. And while LLMs showed promise in enhancing empathy, they also posed risks. Woebot’s team built an LLM-augmented version for controlled clinical studies to evaluate its effectiveness and safety. Initial results were positive, with the hybrid chatbot maintaining appropriate, supportive interactions.
It’s not perfect, but it shows progress.
And I think that really is why I’m a little underwhelmed at where we are. We saw a massive step when Chat-GPT made it into the world. But, with every release of Claude, or Gemini, or OpenAI, we see incremental step changes - rather than the big application that changes everything. OK, maybe I’m expecting too much, but a little uncanny valley from seeing video avatars are not what is going to save us from the biggest problems of today.
I resisted from researching this before writing this piece, but my openAI sidekick told me what I knew already. I asked which are the biggest problems that GenAI is solving. It confirmed my suspicions. Some of the biggest problems are yet to be solved (verbatim copied):
Art and Music Generation: AI creates unique art and music, inspiring new avenues of creativity.
Design and Creativity Tools: AI democratizes design, enabling professional-quality creations without specialized skills.
Speech Recognition and Virtual Assistants: AI-powered assistants like Siri and Alexa enhance daily task convenience through voice commands.
Content Creation Efficiency: Generative AI rapidly produces high-quality content, saving time and resources.
Customer Service Automation: AI chatbots provide instant, accurate responses, improving customer satisfaction.
Language Translation: AI breaks down language barriers, facilitating global communication and collaboration.
Enhanced Security and Fraud Detection: AI improves security by detecting unusual patterns and preventing fraud.
Personalized Education: AI tailors educational content to individual needs, improving learning outcomes.
Medical Research and Drug Discovery: AI accelerates drug discovery, leading to faster treatment and vaccine development.
Climate Change Predictions and Solutions: AI models climate scenarios, aiding in effective environmental protection planning.
Rather that try and solve all of them, let’s stack our chips on one of the big ones? I mean AI customer service actually raises my blood pressure. Let’s save the world first, and then humans?
Stay Curious - and don’t forget to be amazing,
Here are my recommendations for this week:
Part of my workflow on consuming and learning information has been ramped up by AI tools. One that I’m using regularly is TubeonAI which creates summaries of videos and podcasts in seconds. If you need to up your lifelong learning to stay ahead of the game, sign up for free and then use code BOXOFAMAZING for a premium discount) Sign up
Now
The Schools That Are No Longer Teaching Kids to Read Books: 22nd century thinking or mad. Related: When Kids Talk to Machines
You have multiple ‘social identities’ – here’s how to manage them: As social beings, our identities are bound up with different groups – here’s how to juggle all the various hats you wear. Related (but not one for Mrs H to worry about): ‘Why Can’t I Stop Cheating?’ and (another not for Mrs H to be worried about): Computer says yes: how AI is changing our romantic lives
“Awareness” overload: How TikTok’s mental health content goes off the rails
Mental health awareness is more widespread than ever. Some professionals think it may have gone overboard — especially on TikTok.
Does what happens on your iPhone still stay on your iPhone?: Apple’s famous slogan that suggested total privacy is being tested in the age of AI.
The Scourge of Self-Checkout: The technology promised to make shopping easier. It has done the opposite
Next
Apple intelligence and AI maximalism: Apple has showed a bunch of cool ideas for generative AI, but much more, it is pointing to most of the big questions and proposing a different answer - that LLMs are commodity infrastructure, not platforms or products.
AIs are coming for social networks: An app called Butterflies puts a new spin on how we interact with AI. With Meta and others making similar moves, social media is about to get a lot weirder.
Perplexity Is a Bullsh*t Machine: An investigation shows that the AI-powered search startup Forbes has accused of stealing its content is surreptitiously scraping—and making things up out of thin air.
Researchers Say Chatbots ‘Policing’ Each Other Can Correct Some AI Hallucinations: Using a second AI tool as a kind of “truth police” to detect when the primary chatbot is hallucinating. The tool, also a large language model, was able to catch inaccurate AI-generated answers. A third AI then evaluated the “truth police’s” efficacy.
My Memories Are Just Meta's Training Data Now: Meta’s plans to use personal content posted by Facebook and Instagram users to train algorithms suggest our digital histories are being repackaged to teach AI about—and how to mimic—humanity.
Free your newsletters from the inbox: Meco is a distraction-free space for reading newsletters outside the inbox. The app has features designed to supercharge your learnings from your favourite writers. Become a more productive reader and cut out the noise with Meco - try the app today
If you enjoyed this edition of Box of Amazing, please share and help me grow this group. If you share it on LinkedIn, please tag me so I can thank you.
If you are interested in advertising in this newsletter, please see this link