Posted: Thu 21st Sep 2023
Generative AI tools are making waves across the business world. They're helping us to create content, support design work – and even shape customer experiences.
But as we think about integrating these tools into our operations, it's wise to take stock of the potential pitfalls. Let’s look at some of the key considerations here.
The concept of an "AI hallucination" sounds whimsical, but it refers to the times when AI generates information that’s inaccurate. Essentially, when it makes things up.
The problem is that the generated output from tools like ChatGPT can sound very viable and authoritative, so it's easy to miss these "hallucinations". This can lead to misleading or inaccurate content, which is never a good thing for business.
My advice: Fact-check everything before you put it live. It’ll save any issues further down the line.
Generative AI models are trained on huge datasets. Some of this content will contain copyrighted material. As a result, there's a chance, albeit a small one, that your generated content could mirror something that's already been published.
Of course, it's not easy to spot this content, which leaves you open to unintentional copyright infringements or even accusations of plagiarism.
Think about the content you put into AI models
When using AI platforms like ChatGPT and Bard, you get better results when you add more customised information about your project. Most of the time this works well. However, it's important to consider the information you're handing over, as some AI tools will use it for future training.
Earlier this year, Samsung banned its staff from using generative AI across the business for this exact reason. In a report by Bloomberg, employees were said to have accidentally leaked three pieces of confidential information to AI code and transcription tools.
My general rule of thumb: If you wouldn't put your information into a public forum, don't share it with generative AI models.
Ethics and bias
As we know by now, generative AI tools are only as good as the data they're trained on. If the training data has biases (which most of it will), the AI can produce biased results.
This isn't just a tech issue, it's an ethical one. So, be cautious and make sure your AI-generated content aligns with your company values and ethos.
What about the environment?
The environmental impact of AI doesn't get much coverage. But, as you can imagine, carbon emissions and water usage are high.
According to a recent article in The Guardian, researchers estimated that the training of GPT-3, a predecessor of ChatGPT, led to emissions of more than 550 tonnes of carbon dioxide equivalent. This is like flying between New York and San Francisco on a return journey 550 times.
The same report also states that the early iterations of ChatGPT would use 500ml of water – a standard-sized water bottle – in responding to 20 questions.
Researchers are obviously working hard to reduce the processing power of AI, but we are still some way off. So, what can we do now?
We could use AI where it really makes sense and use other tools (and our own brilliant brains) whenever we can.
For example, you can use ChatGPT's new Code Interpreter to create GIFs or generate QR codes. But there are perfectly good non-AI tools that already do the same thing. Consider cutting the processing power and using them instead.
Watch this webinar to find out how you can use Midjourney and Open AI's Dall.e 2 to create images for your blogs, brochures and presentations:
Be careful of over-reliance
Over-reliance is something else to bear in mind. Yes, it's tempting to let AI take the reins, especially when it can give us quick results. But remember, regulations will come. And there are plenty of court cases happening right now, which might change our access to tools like ChatGPT.
There's something else too. Your human imagination and skills are like nothing else. So, when it comes to creating compelling content for your business, don't forget to use your brilliant mind as well.
Keep innovating. Keep thinking
From speeding up content creation to inspiring creativity, and delivering insights and automation, there's no doubt that generative AI has so many advantages for small businesses. It's just important to approach these tools with a solid understanding of their limitations and pitfalls.
By staying fully informed and treading carefully, you can harness the power of this amazing technology, while avoiding any sticky situations in the future.