How can machines and humans coexist harmoniously? Is there any possible way to sort this problem out? If so, how should we implement the solutions correctly? Will machines, namely AI, make us unemployed? And the list goes on. There are a lot of questions raised when it comes to talk about the ethical challenges of artificial intelligence. There are a lot of thoughts that pop into our minds about the ‘broad’ topic of the ethics of AI. But first, let me define artificial intelligence. Some definitions fit in but one of the best definitions that I’ve found is the UNESCO definition:
“A dynamic understanding of AI is, as systems with the ability to process data in a way which resembles intelligent behavior.”
AI ethics, on the other hand, is an interdisciplinary area that searches to seek solutions or suggestions about the problem area created by artificial intelligence and its usage. It is also one of the emerging and urgent fields which need deep discussions and suggestions.
Artificial intelligence has affected our lives in some way. Since the beginning of the 2010s, the topic has gained momentum and it is one of the hot topics, recently. The intelligent personal assistants (IPA), self-driving cars, and AI-generated healthcare systems are some of the main fields that work by AI and are fed by AI. The usage of AI in so many sectors bring us to a point where we need an ‘urgent’ call to talk about the ethical dimension of AI.
Ethics are important when it comes to mentioning artificial intelligence. Different topics are related to AI ethics. According to IBM, companies are experiencing unforeseen consequences in some of their AI applications, due to poor upfront research design and biased datasets. For instance, Amazon once biased potential job candidates by gender for open technical roles, and they had to scrap the project.
On the other hand, as the time went by, AI will dominate some sectors and there is going to be a shift in the demand for jobs which will cause unemployment. Or we can also observe the ‘environmental’ side of AI in terms of ethics because as new AI technologies are introduced there is going to be a need to adjust it to the environment as well.
The topics, or principles, under the title of AI ethics, are transparency, fairness & justice, responsibility & accountability, nonmaleficence, privacy, beneficence, freedom & autonomy, solidarity, sustainability, trust, and dignity (1). However, more principles may be added. Artificial intelligence influences our lives in every way. It affects individuals, society, and the environment mainly. All these principles related to the ethics of AI can be multiplied. You can add psychology, democracy, algorithm abuse, human relations, etc. but the ones I mentioned above are usually counted as the main principles.
Photo by Markus Winkler on Unsplash
The countermeasures related to this issue usually follow a top-down approach. Particularly governments and non-governmental organizations are the pioneers who acted about the ethics of AI. Recently, the European Union has put forth the EU AI Act to prevent the harms of AI on the member states and citizens of the EU. The Act has good points however, there are loopholes related to ethical issues such as the privacy of the citizens, transparency, or accountability of the deployer companies. On the other hand, if it will be endorsed, it is going to be an example for other countries in the world. The non-governmental organizations in Europe and around the world raised their concerns through open letters and campaigns on the EU AI Act and wanted them to address the issues more elaborately.
Also, we see other governments launching their own AI strategies. This is really beneficial in terms of AI ethics because it sets the legal framework agenda for the governments to make themselves ready to cope with the legal and ethical problems of AI. However, according to the ‘Government AI Readiness Index 2023’, scores in the data and infrastructure pillar show an important difference between high and low-income economies, which illustrates an existing global digital divide. According to the index, the lowest region in terms of data availability and infrastructure is Sub-Saharan Africa and the highest region is North America (2).
From an ethical perspective, we cannot make sure AI does not harm people if the country does not have a consolidated democracy and human rights past. The top-bottom implementation of AI ethics through legal acts may prepare the proper ground but we need to dig in more to get a concrete answer.
If we look from a different viewpoint, the bottom-up approaches able us to tackle the ethical problem differently. The bottom-up approach basically allows us to help AI deployers to train AI by getting data from experiences it gets with its interaction with people. But just like the top-bottom approach, it does not give us the room for maneuver. This approach enables AI itself to train as a human child in order to become ethically qualified. I kind of disagree here, because it will bring us to the topic of bias and privacy. Eventually, we can still benefit from the bottom-up approach of AI since it does not toss aside the individual and society part of the ethical process. As Huang and others suggested in their paper, I agree with them to apply hybrid implementations to find a midway that will be more accurate and solid because the hybrid approaches combine the advantages of top-down and bottom-up approaches.
Photo by Andy Kelly on Unsplash
As Basak mentioned in her paper ‘Feel Free To Destroy’ “…And I ask one last question. What if one AI software superior nation feels free to destroy the other, just because their army is full of AI robots dedicated to killing the enemy? Such an apocalyptic but possible scenario…” well, this paper was the thing that pushed me to write something about the ethics of AI. As individuals and daily users of recent technology, we should not turn a blind eye to this topic. Even if the country/company that has the most capable and strongest AI system can dominate the world, with the proper alignment perspective, we can decrease the harm it does to us, society, and the environment. In that kind of situations, I usually remember a quote from the movie ‘Kingdom of Heaven’ (dir. Ridley Scott)
“None of us know our end really or what hand will guide us there. A king may move a man, a father may claim a son, that man can also move himself, and only then does that man truly begin his own game. Remember that howsoever you are played or by whom, your soul is in your keeping alone.''
Even though those who presumed to play you be kings or men of power…” On ethical issues, we have to take action and raise our voices against governments or companies that have maleficence towards society or the environment. We have to go hand in hand with the governments or companies if we can observe that this is going to harm us. We should get run off their feet, so to speak. Thus, the machines and humans coexist harmoniously in that way.
SOURCES
1- C. Huang, Z. Zhang, B. Mao and X. Yao, "An Overview of Artificial Intelligence Ethics," in IEEE Transactions on Artificial Intelligence, vol. 4, no. 4, pp. 799-819, Aug. 2023, doi: 10.1109/TAI.2022.3194503.
keywords: {Artificial intelligence;Ethics;Guidelines;Privacy;Government;Systematics;Security;Artificial intelligence (AI);AI ethics;ethical issue;ethical theory; ethical principle}, https://ieeexplore.ieee.org/document/9844014?denied=