Artificial intelligence is a handy tool but does not free us from thinking
Artificial intelligence offers us new ways to work, create ideas, or be inspired. On the other hand, there is a fear that AI will take over jobs or be used to create and spread false information. Not long ago, quite credible images of Pope Francis in a thick winter coat and the arrest of former US President Donald Trump spread around the world. One thing is certain: artificial intelligence is changing the world in unprecedented ways.
But let's take a few steps back and address the question of what artificial intelligence is. Chat GPT, the conversational bot launched by OpenAI, which is perhaps the hottest topic in AI right now, gives us the following answer:
Artificial intelligence (abbreviated as AI) is a field of computer science that aims to develop systems and algorithms capable of performing tasks that usually require human intelligence. Such tasks may include image or speech recognition, decision-making, creative design, forecasting, and automated control. AI applications are extensive and can cover areas such as healthcare, transportation, economy, marketing, gaming, and cybersecurity.
Let's leave the AI application aside for now and turn to researchers at the University of Vaasa who specialise in AI. Johanna Hautala, associate professor in regional development and innovation policy, studies the processes of creativity and data refinement that utilise AI.
– I am curious but critical towards AI. At the moment, I am a visiting researcher at Stanford University in Silicon Valley, and I have already come across many self-driving cars. I was very excited to share this with my local colleagues, but they were slightly amused because for them self-driving cars are a part of everyday life.
Hautala leads two research projects related to AI. One project, funded by the Academy of Finland, focuses on AI from the perspective of knowledge in Finland, Singapore, and Silicon Valley. The other project, funded by the Kone Foundation, investigates how AI influences creativity in research and art.
– AI fundamentally changes and at the same time questions what knowledge and creativity are. Traditionally, we have had a human-centred understanding of who can have knowledge or be creative. Now, we refine knowledge and are creative together with self-learning AI. Can AI have knowledge? Can AI be creative? So, what are knowledge and creativity in the era of AI? When working with AI, we must question the familiar and create new practices, explains Hautala.
Does AI take our jobs?
AI will significantly change the future of how we work. AI can change certain tasks or even make some jobs redundant. However, new technologies can also create new tasks and jobs. It is important to note that AI will gradually change how we work and different fields will be affected in different ways. Nevertheless, it will also offer many opportunities to develop new ways of working and creating value. (Chat GPT)
Johanna Hautala has described the collaboration between humans and AI as a joint creativity or knowledge refinement process. According to her, AI can play a significant role in certain stages of this process, and it seems that the role of AI grows stronger over time.
– For example, an artist who follows the learning process of AI may discover entirely new perspectives on their subject. AI can suggest, inspire, process, and summarise vast amounts of data. AI can free up time from detailed and repetitive tasks, allowing humans to focus on managing the process and creating ideas.
– An interesting example of this is the play "Love Simulation Eve" at Espoo City Theatre. I cannot even say that I went to see it, as I would usually say about a play, but I went to experience it. The entire traditional theatre performance and the spatial practice of theatre had been reimagined. Instead of being a passive spectator sitting and watching the actors on stage, I was guided by assistants through different spaces and became part of the play in a virtual reality, where I interacted with the characters and influenced the story.
By outsourcing routine tasks to artificial intelligence, you save time and work more efficiently. People can focus on tasks that require problem-solving skills, brainstorming, creativity, or combining ideas. We can only speculate how AI will ultimately change the workplace and the tasks we perform.
Assistant Professor Mikko Ranta, a researcher in text analysis and AI, believes that AI will cause a major revolution in the workplace over the years: some tasks will change, and some may disappear.
– I believe that humans are needed in all creative work, because we want to see a human touch. Chat GPT is a good tool for handling tasks, but it cannot completely replace humans.
Ranta uses Chat GPT almost daily when he wants to quickly convert his thoughts from a list of bullet points into a draft text. The text he gets from Chat GPT will not be completely finished but it is a draft that he can edit, check, and further develop.
Ranta uses language models when analysing texts. The key focus of his research is the use of AI methods for various research questions in accounting. For example, Ranta's research material may consist of company financial statements or job advertisements. With the help of AI, he then finds emerging themes.
Stay alert – AI has its shortcomings
While AI has already made significant progress in many fields, its use also comes with many risks and challenges. The development of AI must be closely monitored and supervised to ensure responsible and ethical use. (Chat GPT)
According to Johanna Hautala, addressing the challenges, risks, and threats related to AI requires various experts, such as AI developers, researchers from different fields, and decision-makers to collaborate.
– If, for example, we were to accept text generated by AI as correct information without questioning, verifying, or justifying it ourselves, we would end up making decisions that are not justified or learning things that are not true.
Hautala emphasises that Finland is an information society where the activities of companies, universities, and public administration are based on well-founded knowledge. According to Hautala, careless use of AI in data refinement processes would jeopardise the entire idea of an information society.
Mikko Ranta explains that language models like Chat GPT are typically trained by feeding them a massive amount of text snippets taken from the internet, parts of which are masked. The system must learn to predict what would be an appropriate text for the masked parts. By practising with a sufficiently large text dataset, the model learns to engage in surprisingly good conversations with users.
– When people play around with Chat GPT, they may think that the model is smarter than it is. It would be good to understand that the application is not as intelligent as it appears, and it also has its shortcomings.
When presented with a question, Chat GPT uses the enormous amount of text it has been fed as its source. Based on that, it gathers and combines information to deduce a sensible answer.
– So, the application is not smarter than that. It does produce a significant amount of text in an instant and seemingly sensible answers to any questions, saving us humans time and effort. However, it cannot know anything beyond what is available on the internet, which is one of the challenges because any nonsense can be written online.
If Chat GPT does not have the answer to a question, it comes up with things on its own. In the era of AI, people need to be more media literate than ever to distinguish reliable information from unreliable and true from false.
– Once when the application came up with things that were not true in response to my question, I asked why it was lying. Chat GPT apologised and offered a new suggestion. So, it is important to realise that if you ask Chat GPT to search for information online, the responsibility for checking and verifying the information always lies with you.
Chat GPT has become a popular phenomenon in an instant, and it is a topic that is raised both in the workplace and in the media.
– AI is now really starting to be useful. It has been researched for almost a century, but now there is finally enough data to train AI models to be sufficiently reliable, explains Ranta.
Ranta states that when an AI application like Chat GPT becomes available to the public, it has a huge impact. Therefore, it is extremely important to study it and its effects thoroughly.