AI - a popular buzzword that was up until now mostly a fun gimmick, or a topic of discussion on TV or talk shows. But that time has passed. Those AI foundation models have evolved so greatly that this year they will quickly find their way into the business world. AI has become more advanced, more accessible, and available in more applications than ever before. In this article, we summed up the biggest AI trends you should look out for.
GPT, the language model everybody talks about
One of the most notable AI improvements this year will be GPT. But what is it and how does it work? GPT stands for Generative Pre-trained Transformer and is a machine-learning model that uses data to generate text. The 3rd version of its kind has 175 billion parameters and has been fed around 570GB of text information gathered by crawling the internet. In human language, this means that this is the first time that a language model answers you with such high quality that it is difficult to distinguish whether the output was written by a human or GPT3.
But how cool this may sound, it also has its limitations. Its last data training was in September 2021 which means that if you ask chat GPT about any event that happened in the past year, it will either be unable to respond or will provide you with outdated information. Secondly, due to its growing popularity, the website is sometimes overloaded and inaccessible. However, a payment model will eventually resolve this.
Finally, biased data used to train the model can potentially cause AI systems to be inaccurate. For example, if a model is trained on data that is disproportionately skewed toward one particular demographic, it may perform poorly when applied to other groups. Therefore, you should always be critical of the responses generated by chat GPT.
Hopever OpenAI is working tiredlessly on imporiving this technology, and other competitors such as Google have already announced that they will release similar AI innovations in the coming answer.
AI as a content creator, report generator, or your own personal assistant
Another exciting aspect of AI this year is its accessibility. You will probably have noticed that the news and your social media have been flooded with AI-drawn paintings bij DALL·E 2, ChatGPT-generated content, or friends talking about it. AI isn’t a vague concept anymore. It is here, and it is available to both you and me, too big and small companies. Although GPT is now only accessible through an often overloaded web page, an API will soon as well be released, allowing GPT to be integrated into any enterprise solution. The possibilities of this are going to be tremendous.
Microsoft, for instance, has invested billions in OpenAI, the organization behind ChatGPT, to deploy its solutions in their office tools. But Microsoft isn’t the only one to collaborate with OpenAI. Copilot, GitHub’s AI pair programming tool, is also gathering forces with ChatGPT. Together they create a "pair programmer", an AI assistant that supports programmers with tiresome daily tasks so that they can concentrate on important matters. Within Rollo these technologies are already adding tremendous value to our day-to-day job an we expect this to increase even more with the upcoming improvements.
Furthermore, Marketing tools like StoryChief, a social media scheduling tool, have enrolled AI in their software to write catchy social media copy for you. And also within smaller SMEs computer vision models, optimization models and prediction models are becoming easier and more accessible thanks to improved software tools and cloud services. ChatGPT can, for example, be trained on a company's data to generate reports in natural language. Employees can make better decisions and save time and resources doing this. For example, Rollo has just developped a GPT based language model that helps teachers write personalized feedback to their students.
AI for Science
Not only is AI gaining ground in the business world, but a huge AI community is also rising in the science world as well. Last year Alphafold and Alphatensor, 2 AI models showed some pioneering results of which we have only seen the beginning. AlphaFold predicts a protein’s 3D structure from its amino acid sequence. This will help us understand the structure of proteins and advance biological research. And Alphatensor can find efficient and fast matrix multiplication algorithms. This more efficient way of matrix multiplication will not only help the scientific computing field but even general computing performance (of smart phones, PC's etc.) will be impacted significantly.
What about GDPR and regulations?
Finally, we can expect to see a large focus on ethics and transparency in AI. As the use of AI technology becomes more widespread, there will be a growing need for clear guidelines and regulations to ensure that it is used responsibly and ethically. The EU is already working on a GDPR for AI, the so called AI Act. Overall, while AI has the potential to be incredibly powerful and useful, it is important to be aware of the potential biases and to use the technology ethically. The impact innovations such as ChatGPT will have on our society will be enormous. Early government regulation is therefore necessary and Rollo is happy to support these efforts.
In conclusion, there are many fascinating advances in AI to watch out for, ranging from the development of AI for science to the widespread application of machine learning models in the business world.
It is clear that AI will become even more accessible to the general audience and small and medium size companies in the coming year.
Wondering what AI might entail for your business? Feel free to drop by and have a chat. We serve the best coffee in town. ;-)