Asimov’s short story Runaround (1942) centers around the “three laws of robotics,” which center around principles that robots must follow in the interest of welfare for both humans and the robots themselves. Today, these laws are relevant in discussing AI ethics (Britannica, 2024). Ethical concerns inevitably arise as AI becomes more commonplace in fields ranging from education to art to medicine. So, what ethical issues arise with AI, and why are they important?
What is AI Ethics?
AI ethics studies how to optimize AI's beneficial impact while reducing risks and adverse outcomes (IBM, n.d.). This includes setting regulations for AI training and usage and preventing undesirable outcomes that contradict human morals, such as bias and abuse. In creating and using AI, many potential problems could arise that need to be resolved or regulated, and they will be discussed below.
Copyright & Ownership
By current standards, whoever created a work, such as a news article or drawing, owns it. Plagiarism and unauthorized use are considered unlawful and unethical. This raises the question of whether AI must follow similar moral codes when using the works of others. Generative AI, such as the large language model (LLM), must use massive amounts of existing work as training data. Obtaining usage rights for millions of text excerpts and images online is undoubtedly unrealistic so that developers may skip this step.
However, when the AI produces notable content similar to existing sources used as training data, it can be considered plagiarism and upset the original creator of the source.
Many also argue that AI does not create the content it produces, as they consider it a mindless amalgamation of its training data. Hence, AI-generated content would not be owned by the AI model or the person who used it.
Autonomy & Responsibility
AI at present already helps make many important decisions, from hiring someone for a job to determining whether a suspicious mole is skin cancer. In these decisions, various levels of autonomy and power can be given to AI, but they come with various risks as this level increases.
If an AI with some degree of autonomy makes a mistake, who is responsible for it (Müller, 2023)? Currently, there is no universal legislation that determines this. However, researchers and ethicists are working to fill the gaps by developing ethical frameworks to follow (IBM, n.d.).
Bias
As the saying goes, “Garbage in, garbage out”. When making decisions, AI tends to follow trends in its training data. However, when the data is a historical record of past decisions, they may contain unwanted biases. This can cause the AI to make judgments based on irrelevant factors, including discriminatory preconceptions about group members.
Such historical bias was discovered in a recruitment screening system at Amazon (discontinued in early 2017) that discriminated against women—presumably due to the company’s history of discriminating against women in the hiring process (Müller, 2023). Such biases are commonplace in employment and legal settings, hindering societal progress toward equality and causing unwanted harm to many.
Conclusion
Artificial Intelligence is still new to the world but has been growing exponentially. Due to its nature, its usage can cause harmful side effects, but a common consensus on the ethical way to use AI has yet to be established. As AI becomes deeply rooted in our daily lives, we humans will likely have to shift how we think about morals and ethics to accommodate technological progress.
Reference List
MedTechUCL (2022). The Ethics of Artificial Intelligence (AI). [online] MedTech UCL. Available at: https://uclmed.tech/the-ethics-of-artificial-intelligence-ai/ [Accessed 31 May 2024].
Müller, V.C. (2020). Ethics of Artificial Intelligence and Robotics. Winter 2020 ed. [online] Stanford Encyclopedia of Philosophy. Available at: https://plato.stanford.edu/entries/ethics-ai/#ArtiMoraAgen [Accessed 31 May 2024].
Reference listIBM (2023). AI Ethics | IBM. [online] www.ibm.com. Available at: https://www.ibm.com/topics/ai-ethics [Accessed 31 May 2024].
Tikkanen, A. (2022). Three Laws of Robotics | concept by Asimov | Britannica. [online] www.britannica.com. Available at: https://www.britannica.com/topic/Three-Laws-of-Robotics [Accessed 31 May 2024].
コメント