Some Potential Concerns with AI
i) human challenges (see graph below)
(source: Prosci, 2024b)
This is based around a lack of understanding about AI and people have
"...expressed concerns about over-reliance on AI leading to decreased human judgement and loss of critical thinking skills, as well as potential to perpetuate errors or bias that is present in the data. Using AI without applying human judgement is particularly concerning because some GenAI systems can and do generate incorrect responses or 'hallucinate' without clear signals of the accuracy. To avoid significant problems, AI users need to develop critical skills to verify information, identify inaccuracies and edit false information..."
Prosci, 2024b)
ii) resistance to AI (resistance is a natural reaction to change and can take many different forms, eg disengagement to sabotage, and result in loss of the important staff.
"...Resistance to AI is often based on the fear of the unknown or fear of failure. Is the chatbot going to replace customer support as a function? Will the team of data analysts lose their jobs because their work can be automated as part of AI systems? Will I have the skills I need to perform effectively the new reality?..."
Prosci, 2024b)
Need to uncover the root causes of resistance and address them; need to
"...reframe them as opportunities for people to build new skills, refreshed roles, and grow through using AI..."
Prosci, 2024b
iii) integrating technical aspects of AI with the human factors that influence adoption and usage of change so that its use is effective
iv) ethical issues around using other people's intellectual property (IP) (need to acknowledge and recognise when you are using other people's intellectual property that GenAI generates)
Some current (2025) limitations of AI include
- unable to determine the accuracy of information and data
- unable to handle infographics (this is improving)
- despite its strength in analysing the past and present, it has limited potential to forecast and/or make predictions about the future
- unable to think creatively
- unable to show the full range of human emotions.
Resource Usage with AI
As AI uses considerable amounts of electricity, this is its greatest challenge: having the computing infrastructure including data centres adequately supplied with electricity.
"...The electricity consumed by data centres globally will more than double by 2026 to more than 1000 terawatts hours, an amount roughly equivalent to what Japan consumes annually.
Updated regulations and technological improvements, including efficiency, will be critical to moderate the energy consumption from data centres..."
International Energy Agency as quoted by Camilla Hodgson, 2024
In addition to power, large volumes of water are required to cool the data centres.