Artificial intelligence has been making the news lately. It is one of those buzzwords that make people feel like they need to know about it, even if they don’t know what it is says Peter DeCaprio.
AI has become so popular recently that everyone wants in on the action. While many companies do not actually use real AI technology, they include “AI” or “artificial intelligence” because it makes for great marketing material. The problem is that sometimes these claims aren’t true at all.
Here are 7 unforgivable sins artificial intelligence by fake AI claims.
1. Calling a chatbot “intelligent” just because it can answer questions.
Chatbots have become a trendy way for companies to take care of basic customer service requests or deal with website inquiries quickly and cheaply, but these bots are far from intelligent. They generally only know how to respond based on keywords within the question, and they cannot hold conversations as humans can. Peter DeCaprio says this limits their ability to help customers in any meaningful way, which is why many companies that use them set strict time limits or courses of action for users.
Many businesses use speech-to-technology because it saves time and makes it easier to communicate with customers and leads explains Peter DeCaprio. For example, a customer service agent can use voice recognition software to quickly type what a customer is saying instead of having to input everything through a keyboard.
The problem here is that the security and privacy policies for this data are rarely made clear to users. While businesses are legally required to disclose this information, they often hide it deep within their terms of service or other documents. Even if transparency is present, there are legitimate questions about whether or not users understand how much data they are giving up when using these services.
3. Claiming your chatbot will improve over time without mentioning that you’re teaching it based on human interactions.
This one ties together with number 1 – because the bot only understands what you tell it based on keywords, each interaction that a business has with a customer is actually recorded and used to improve the bot’s future answers. This means that customers will have to repeat themselves over and over again, repeating their own answers to questions so the chatbot can learn from them.
4. Advertising deep learning or neural network technology without explaining what it does or how it works.
If you’ve been reading up on artificial intelligence, then you’ve probably heard of deep learning. It’s a type of AI that uses advanced algorithms to process information about objects, images, language, sounds, and more in order to generate insights and make predictions. Neural networks are part of the technical side behind deep learning – they’re systems that use low-level data, like images and text, to interpret high-level meaning.
5. Using big data as an excuse for not giving personal attention or ignoring the limits of predictive analytics.
The age of big data has come with many benefits, such as greater insight into customers and their habits. But this also means that businesses now have access to far more information than ever before – and sometimes it can be overwhelming.
Companies that want to use AI in a smart way need to realize where they’re at, what information is available, and how much they can do with it says Peter DeCaprio. This includes understanding the limits of predictive analytics (more on those here). Using human intelligence when appropriate and incorporating feedback from customers. So you know which services are working well and which need improvement.
6. Claiming your product will predict the future without explaining how it works behind-the-scenes.
We all love to know what’s going to happen in the future. This is why companies that claim their AI can predict things are so appealing. Unfortunately, this promise often comes with big caveats. Especially when it comes to products that use deep learning or neural networks. Because of how complex these types of technology are, many claims. That they can predict the future ends up being little more than false advertising. Which means you could be wasting your money. The best way for companies to avoid this issue is by explaining exactly how their software works. And what data they need from customers before making any predictions.
7. Promising to build a “fully intelligent” AI without explaining what it will do.
This is similar to number 1 – and like that sin, it’s something that can trip up companies big and small. When you’re developing your own chatbot, you might hear this idea of building an intelligent bot repeated. But what does that actually mean? What problem are you solving by making a bot more intelligent? It could be anything from enabling the bot to answer specific questions (like question 3) to solve complex problems (such as number 2). Without knowing exactly what intelligence looks like for your company. It’s hard to know how far along you actually are in development.
A chatbot might seem like a simple piece of technology. But it’s actually a very complex piece of programming that requires interaction with customers. To record data and improve the bot’s performance says Peter DeCaprio. If you’re thinking about using artificial intelligence for your business. Make sure you understand how it works and what it can do.