OPINION - Finbarr Murphy: I’m glad that we’re beginning to see some push back on the idea that AI is a silver bullet to solve every problem. The reality is, it’s just another tool in the toolbox. It’s true that in some situations it can make a significant impact but sometimes there are other tools more suited to the job. A paintbrush is perfect for creating a stunning piece of art but when it comes to painting a wall, a roller will do the job in twice the speed with a better finish.
Also, as the article in the The Guardian argues, the promise of AI can be a welcome distraction from solving the more pressing issues of today. Is AI really a bionic duckweed? I’m not opening that particular can of worms today; however, to use the healthcare example in the article, without the right groundwork and addressing the systemic issues plaguing the NHS, the impact it can have will be minimal and contained.
If we want to realise the true potential of AI, we must understand if we’re ‘data mature’ enough to benefit from it and assess it against the other tools available. Instead of AI for AI’s sake – we should be balancing innovation with what will make the biggest impact right now.
Sorry for the shameless plug, but this is exactly what we’re doing at Modular Data Helping you understand your level of data maturity, giving you a roadmap to get to where you want to be, and selecting the right tools to help you get there – whether that’s AI or not.
https://bit.ly/4cBZt0k
#datainhealthcare #AIinhealthcare #dataproducts #datamaturity #maturityassessment
Partner, Enterprise Data Strategy & Activation at AMP42
1moI love this topic, Shweta! I worked on building a model for class to predict the timing and magnitude of volcanic eruptions. The trouble with this is exactly the problem you discuss. Since volcanic eruption data require highly sensitive instruments in harsh environmental conditions and the relative frequency of actual eruptions is low, there isn't much of this data available. So, a team of researchers came up with a mathematical model to generate realistic eruption data based on existing environmental conditions such as conduit radius, magma density/viscosity, atmospheric pressure, etc. It was essential to have enough data to train and test our model, and it is clear to me how promising synthetic data are when real-world data are scarce. 🌋 👏