Dear People,
I often go to my local cupcake shop to buy my favourite Raspberry Cupcake. The owner's cupcakes are great, but he usually has many unbought cupcakes sitting around, and some flavours are frequently out of stock.
Watching him operate, I get excited because selling cupcakes generates data that could be invaluable if he had access to AI. AI systems are excellent at spotting patterns.
An AI could help the owner predict which cupcakes sell well on certain days, improving inventory management and increasing his revenue.
You might say, "Hey, Akanksha, this is a small cupcake shop. What's the big deal?" To the cupcake shop owner, improving revenue by a few thousand pounds yearly is a huge deal. There's a lot of hype about AI's need for massive data sets, and while more data helps, AI can work just fine with modest amounts of data, such as that from a single cupcake store.
The real problem is that a small cupcake shop store could never justify the cost of hiring an AI team.
In the United Kingdom, there are about half a million independent restaurants. Collectively, they serve tens of millions of customers, but each restaurant is unique, with different menus, customers, and sales recording methods. No one-size-fits-all AI works for all of them. What if we could enable small businesses to use AI?
AI for Small Businesses
Consider a company that makes and sells T-shirts. An accountant there could use AI for demand forecasting, determining what funny memes to print on T-shirts by analysing media trends. A store manager could use AI to recommend product placements to improve sales. A buyer could use AI to decide whether to purchase fabric at a specific price or wait for a better deal. Quality inspectors could use AI to scan fabric for defects, ensuring high-quality products.
Today, large tech companies use AI to solve these problems effectively. However, a typical T-shirt maker, auto mechanic, retailer, school, or local farm uses AI for none of these applications. Each T-shirt maker is different, so no AI system works for all. Large companies like pharmaceutical firms, car makers, and hospitals struggle with this issue, even in other industries.
The Long-Tail Problem of AI
This is the long-tail problem of AI. Suppose you were to list all current and potential AI projects in decreasing order of value. In that case, you'd see a few precious projects (like ad targeting or web search engines) and many unique projects with significant aggregate value but requiring custom solutions.
Enabling Everyone to Build AI Systems
How can we enable small businesses and individuals to build AI systems that matter to them? Creating an AI system required writing extensive code for most of the last few decades. While learning to code is becoming more accessible, not everyone has the time for it. But there's an emerging new way to build AI systems that lets more people participate.
Just as pen and paper—a superior technology to stone tablets and chisels—were instrumental in widespread literacy, new AI development platforms shifted the focus from writing code to providing data.
This is much easier for many people. Several companies are working on such platforms. For instance, an inspector wanting AI to detect fabric defects could take pictures, mark areas with defects, and train the AI with this data.
By adjusting the data provided to the AI, users can help the AI get smarter. An inspector could build a custom AI system to detect defects in all fabric used in a factory within a few hours to a few days. This technology can similarly empower bakers, farmers, and furniture makers to improve the quality of their products.
While these platforms still need a few more years to become user-friendly for every small business owner, many are already exceptionally helpful with some training. This means we can empower every accountant, store manager, buyer, and quality inspector to build their own AI systems.
I hope small business owners, like the local cupcake shop, will use this technology. AI is creating tremendous wealth, and by democratising AI, we can ensure this wealth is spread far and wide across society. Hundreds of years ago, few understood the impact of widespread literacy.
Today, only some understand the effects of democratidemocratisingAI. In the coming era, we'll empower everyone to build AI systems for themselves, leading to an inspiring future.
Thank you for reading.
Akanksha
Comments