Many people point out the benefits AI can bring to the field of cybersecurity. It has a good track record dealing with malware, and it can help detect weak passwords, shoring up our defense against some types of cybercrime. At the same time, AI can be manipulated by cybercriminals to steal information, or even biometric data, decreasing the privacy of individuals. There is also the possibility of malicious code generated by manipulated AI being put to use by unknowing individuals. Watch the video below for more information.
According to Adam Zewe at MIT News, "the computational power required to train generative AI models that often have billions of parameters, such as OpenAI’s GPT-4, can demand a staggering amount of electricity, which leads to increased carbon dioxide emissions and pressures on the electric grid." Additionally, these systems need frequent cooling, which can use a lot of water, potentially interrupting local ecosystems. These environmental impacts are real and an increasing problem. Some experts believe that these impacts may be mitigated by the possible solutions to climate problems that generative AI might identify, but this possibility has yet to be actualized.
On this page, we will explore some of the ethical issues surrounding AI use. We will focus on:
Bias creeps into AI several ways. First, what AI learns includes the existing biases of its dataset. So, as the video below points out, if you develop AI that will determine who gets a loan, and you feed it the profile of who was previously successful at obtaining a loan, it inherits the existing biases of that process. Second, most people do not train generative AI, but exist under its influence nonetheless. Who gets the power to train the AI can influence what type of perspectives are represented. More women and minorities need to participate in developing AI. The more diverse the perspectives are that train AI, the more accurate the models will be. Third, not everyone has equal access to using AI.
There are three major ethical issues surrounding AI and intellectual property. These are described in the Wall Street Journal video below. They are:
All of these issues are currently under litigation; because we don't know the outcome of these cases, we cannot be certain whether the content we create using AI does not violate copyright law. Therefore, using AI data can be ethically problematic.
Just as the Industrial Revolution impacted manual labor, generative AI is revolutionizing the workforce. As the influence and use of AI at the workplace increases, workers are both excited and fearful over its possible impact. According to Goldman Sachs, roughly two-thirds of workers will be impacted by AI automation and AI could substitute for about one quarter of current jobs. What work we do, and what jobs we train for may change rapidly over the next several years. It is important to be planning for these possibilities at the individual and institutional level.