Artificial intelligence (AI) can feel like an ever-present topic of conversation at the moment. This is good, as we must keep having conversations about it. The development landscape is moving rapidly and unless we ask some questions and take decisive action, unethical usage of these tools could bloom out of control.
Regulations and accountability have to be part of the conversation we’re having. Protocols need to be in place to give us time to consider how tools should be used. While legislation may come eventually, many of the decisions need to be made on smaller scales. Industries, businesses, and stakeholders alike will need to collaborate on regulations and accountability measures that leverage AI in responsible ways while safeguarding against possible problems.
Considering Usage Cases
It can be tempting to be overconfident in AI and grab hold of tools that make operations more efficient. Yet, one of the roles of regulations and accountability is to ensure businesses aren’t using AI for tasks that present ethical risks.
Some of the areas that can be appropriate for mindful regulatory restrictions on use might include:
News and research papers
The public relies on accurate information from news and research stories to make informed choices about their everyday lives. While producing or disseminating texts by AI might seem efficient, many such platforms still have significant issues surrounding the accuracy of generated text. While this may change over time, enforcing regulations on using generative AI for papers and news articles is key to minimizing the public’s exposure to harmful misinformation.
Medical treatment
There are a lot of really positive potential uses for AI to improve patient care, including condition diagnosis and developing treatment plans. As Joyce University notes, healthcare’s new approach to data-based tech can do a lot for both patients and providers. However, these tools require vast amounts of data, often related to patients with similar conditions. This may mean that sensitive patient data is accessed and shared across machine learning systems, increasing its vulnerability to breaches.
The Health Insurance Portability and Accountability Act (HIPAA) provides some regulatory guidelines that apply to the use of AI data collection. Developers and facilities must consider whether their AI usage unnecessarily presents risks of breaches.
This doesn’t mean that AI shouldn’t be used at all on these types of tasks. Rather, it’s a case of avoiding AI’s complete autonomy. Regulations that insist on and outline stringent methods of human oversight and supervision can both minimize ethical issues and identify representatives who are accountable in the event of breaches.
Addressing the Labor Impact
One of the most prominent ethical considerations surrounding AI at the moment is the potential to replace jobs. There are certain professions less likely to be replaced by AI. These are usually skilled trades positions, such as mechanics, carpenters, and electricians that AI tools aren’t advanced enough to replicate yet. Similarly, roles that require high levels of emotional connections – like healthcare workers – aren’t particularly suited to AI.
This still leaves a lot of other jobs that are at risk of being supplanted by AI tools, though. We’re already seeing software that can handle certain customer service tasks, data entry, and even accounting. If companies and industries choose to replace workers without ethical forethought, the outcome could be significant unemployment that our social welfare systems aren’t robust enough to support. This means that unregulated AI use can result in worsening the socioeconomic disparities that are already serious.
Government regulation to restrict job losses due to AI is not exactly forthcoming at the moment. Indeed, while the White House Executive Order has proposed some legislation, this is largely directed toward issues like privacy and security, rather than mitigating job losses. Therefore it is down to individual companies and industry bodies to compel employers to slow down AI adoption that could result in unemployment.
A good example of this is the provisions the Writer’s Guild of America (WGA) secured in negotiations with studios in 2023. This includes prohibiting studios from using AI to edit or rewrite scripts that have been produced by a writer and ensuring AI-generated materials can’t be used as source materials to base scripts on. These may seem like relatively small measures, but they effectively save jobs. They also provide systems of accountability in the event of unethical breaches.
Mitigating Biases
AI systems are programmed and developed by people, with all their flaws and individual perspectives. As a result, the results can be subject to the biases of AI programmers and the information it learns from. This presents an ethical challenge.
Protocols to address gender equity in AI development are essential. Women have historically been underrepresented in science, technology, engineering, and math (STEM) fields. Developers adopting recruitment programs that attract and retain women – alongside professionals from other marginalized populations – at all levels of seniority means AI tools benefit from various perspectives. Tools may be less subject to biases in programming that may result in less ethical outcomes.
Such programs begin with industry leaders and startups taking steps to encourage more diverse contributors to enter STEM education specializing in AI. Representatives need to visit schools, clarifying that there are roles available for people from all backgrounds. Companies can contribute to scholarship funds that are directed toward supporting marginalized learners. Beyond this, individual employers can implement hiring protocols that consider people with non-university educations, as this can lead to onboarding talented professionals from varied socioeconomic backgrounds. By making equity the basic standard for hiring in AI, we can mitigate bias.
Conclusion
Implementing regulations and accountability protocols for the use of AI is key to keeping our use of this technology ethical. This should include restrictions on what the tech is used for alongside standards to mitigate bias, among others. Ethical decision-making is not something we can afford to be slow about, either. By taking action now, our society can find ways to benefit rather than suffer from a technology that is already outpacing legal frameworks.
##
ABOUT THE AUTHOR
Ainsley Lawrence is a freelance writer who lives in the Northwest region of the United States. She has a particular interest in covering topics related to UX design, cybersecurity, and robotics. When not writing, her free time is spent reading and researching to learn more about her cultural and environmental surroundings. You can follow her on Twitter @AinsleyLawrenc3.
