Leah HennigThe bad reputation surrounding artificial intelligence (AI) among Gen Z should not stop the University of Alberta from pursuing AI research. Depending on which side you stand on, “AI research” may sound intimidating. The U of A’s placing eighth in the world and second in Canada as a leader in AI research. That includes research on the development of AI, as well as research that utilizes the technology. It is imperative that the university continues to pursue AI research despite popular ecological and artistic concerns. No one should be preventing potentially lifesaving technologies in scientific research and medical fields from becoming real.
There are two main arguments against the further development and adoption of AI — the first being that the continued building of AI data centres is harmful to the environment. It is a well-argued point that data centres consume immense amounts of water. Data centres also heavily depend on natural resources while emitting many pollutants. A ChatGPT message, on average, uses up to 10 times the electricity that a Google search does, and making the computer that processes it requires up to 800 kilograms of raw materials. Plans for continued construction of these data centres threaten damage to the landscape’s environmental integrity.
The training of generative AI is the other main issue with development in AI. Currently, this is the leading consumer use for AI models. Training involves the scraping of artwork (which is publicly available, though potentially copyrighted) from the internet and combining it in a virtual cesspit, from which the model will learn to produce its own work. Artists have hotly condemned this method of training, as generative AI tends to appropriate art styles at the expense of artists’ opportunities in industries such as film and advertising.
It is not hard to see, through both environmental and artistic factors, why the contemporary Gen Z attitude harbours an anti-AI sentiment. As shown by these irrefutable problems with AI as a technology, why should the U of A continue to pour money and resources into this field? It is necessary to appreciate the nuance of the situation and the environmental and artistic dilemmas. But great change can still come from the use of AI in scientific disciplines.
A paper published in Nature at the beginning of this year outlined the value of AI in research across disciplines. It found that scientists who implement AI augmentation in research triple their number of papers published but decrease the thematic scope, suggesting that AI is better utilized in areas with more established data than burgeoning scientific frontiers. This showcases AI’s situational value. AI may excel in specific areas which involve mass computing, such as protein folding, genetics, or aiding in some menial steps in information synthesis for review articles.
Research into protein folding resulted in one of the most significant advances in modern scientific history — the AlphaFold protein structure database. The 2024 Nobel Prize-winning program is completely free; any user is able to input a sequence of any amino acids (such as protein building blocks) and have AI rapidly predict a probable folding pattern for the resulting protein. Protein structure is vital in understanding mechanisms of action for drugs, gene mutations that result in disorders (particularly diseases like Alzheimer’s and Parkinson’s), and even perhaps the designing of proteins for custom tasks. AlphaFold will enable novel research on these topics to proceed far faster and more thoroughly.
Cleveland Clinic also notes that, in the health-care field, specially trained AI models can lead to more confident diagnoses for imaging technologies like X-rays and MRIs. These uses of AI differ in many ways from the generative AI described above — these smaller-scale models do not require mass data centres, are not trained on stolen artwork, and do not replace human jobs. AI research is a rapidly expanding field which has the end goal of enabling human researchers to increase efficiency and pave new pathways for discovery.
The conclusion to be drawn is one of caution; AI holds much potential, but it is important for our generation to distinguish the capabilities of scientific, specialized AI as opposed to more widely used generative AI. The best solution is to reframe the focus from the production of commercial AI data centres to research, and encourage training models which have direct applications in the scientific or medical fields. The U of A should decisively continue the push toward AI research, as, under the right direction, AI research yields a net positive effect on the world.
