Tools like GitHub Copilot assisting you with coding, ChatGPT serving as a tutor or DALL-E boosting creative projects (actually the pictures for this blog post were created using AI) can often be invaluable. Even though I have some experience using Gen AI tools, I would not call myself an expert as there are still many useful ways of implementing this technology that I have not yet discovered. Gen AI often proved to be a great brainstorming partner or a companion helping explore uncharted territories of new information whether at work or university. However, there are significant challenges I came across, with intellectual property (IP) and accuracy.
Gen AI is an amazing tool if used wisely. Nonetheless, I came across a fair deal of issues when implementing Gen AI into my tasks. The biggest difficulty I had was dealing with intellectual property such as legacy code. Do not get me wrong, Gen AI proves to be a great aid with most coding assignments, however, there are points at which it is close to useless. Developing code that is working based on legacy code is one of these examples. This is one of the biggest concerns in adopting AI at businesses, the concerns about intellectual property. In a case like this one cannot just simply copy and paste the legacy code into an AI assistant due to the obvious breach of confidentiality. In addition, AI models can reuse said code for other users increasing the number of questions around sharing IP with these models.
I also sourced Gen AI for help with university tasks. I often asked ChatGPT to help explain a difficult topic or to help work on new ideas, however, oftentimes I run into issues when asking Chat to provide sources for the information it’s giving me. I thought it would be a clever workaround to fact-check the information provided by the model, however, it turns out that these models have a tendency to fabricate citations. I was a bit shellshocked when I found out but again this was a lesson on the need to be incredibly cautious and aware while using the assistance of AI.
In conclusion, GenAI is a force changing the world at a rapid pace and it would be unwise to not use it. However, the limitations and reliability issues must be kept in mind. I do believe that this situation will only improve, and I am looking forward to broadening the range of AI products I am familiar with but for now, I will stick to using AI as a supplementary tool to the work I do.
