Risks of Using AI in A Business

Since the arrival of ChatGPT and DALL-E 2 last year a whole host of AI tools have come on the market. Some are good, many are so new that they just can’t be good. And they aren’t. Implementing one or a few of these AI tools can be of great benefit. But as with any new technology, there are risks.

In this post, we cover the risks and what to be aware of when integrating Artificial Intelligence into your business, especially small and medium sized businesses. Our next post will look at how to go about getting the most out of AI tools.

Risks of Integrating Artificial Intelligence Tools Into A Business

False Claims: For AI tools to be anywhere near truly useful, they have to be trained on very, very large amounts of data. They also take time to train and learn, especially any that involve Neural Networks or Large Language Models with Machine Learning. If you’re being pitched a tool and it’s new, it’s likely either connecting to another more established tool or is just bunk.

Fake Facts: Text AI tools like ChatGPT are quite good. Sometimes, they’re awful. They have a tendency to make up their own facts and the grammar can be sketchy at times. They need editing more often than not. Google launched its Bard AI product to support search, the result it showed was factually wrong. If Google can make such a mistake, well…

Copyright: Other risks are using images and some text where you could face copyright infringement lawsuits. None of these new GAI tools are creating anything new. They’re regurgitating what is already out there. Fabulous for summarizing complex ideas and arguments and helpful for research. But beware what you put in the public domain.

Marketing Content Overload: One of the first business uses for ChatGPT we saw was creating all kinds of marketing content for corporate blogs. The idea being to get to the top of search pages faster. This is risky. Search engines are detecting AI written content and you could be penalized for this. So you end up losing your position.

Email Sales: Our team has collectively received 142 sales emails written by AI tools. Only 9 of them made any sense. The rest had bad grammar or illogical paragraphs. AI in sales automation so far isn’t very good. Spray and pray is not a sales tactic, especially with horrible grammar. Sales is a relationship system. Be careful how and where you use these tools. We’ve evaluated several, none are very good yet.

Innovation: Again, these Generative AI (GAI) tools are good at distilling complex ideas and making summarisations. They can assist with coming up with innovations, but only as a starting point. It also depends where you’re looking to innovate. GAI is terrible with math, don’t use it for new profit models. Don’t rely on it product innovations and if you’re going to use several innovation tools across the business, GAI won’t help at all.

While there are some excellent ways to use AI in a business, it’s good to be aware of the risks first. It’s especially important to evaluate any claims that are too good to be true. Don’t get caught up in the hype. Experiment first and be aware of the potential impacts on workplace culture.