Real Concerns Arise as AI Technology Drives Discrimination, Deep-Fakes, and Wage Theft
In the ongoing debates surrounding generative artificial intelligence (AI), discussions often revolve around the potential threats posed by AI to humanity. However, while the notion of AI causing the extinction of the human race grabs headlines, it is important to focus on the more realistic and pressing concerns associated with this technology.
A recent statement by the Center for AI Safety, supported by prominent researchers and figures, emphasized the need for global prioritization in mitigating the risks associated with AI, comparing them to societal-scale threats like pandemics and nuclear war. While this attempt to draw attention to the doomsday scenario surrounding AI is effective, it overshadows the actual real-world threats posed by AI, as argued by Emily Bender and Alex Hanna.
According to Bender and Hanna, the significant dangers of existing AI tools lie in wrongful arrests, the expansion of surveillance, defamation, and the creation of deep-fake pornography. These threats are already present in society, unlike the hypothetical risks to humanity’s survival. AI technology, used by many companies, currently enables discrimination in areas such as housing, criminal justice, and healthcare. It also facilitates the spread of hate speech and misinformation in languages other than English. Moreover, algorithmic management systems subject workers to wage theft regularly, and these programs are increasingly pervasive.
Bender and Hanna provide an example of a different real threat by highlighting the ongoing actors’ and writers’ strike. They shed light on how industry moguls aim to replace actors with AI counterparts at a fraction of the cost, while hiring writers on an ad hoc basis to revise the incoherent scripts generated by AI. This demonstrates the tangible consequences of AI encroaching on jobs within the entertainment industry.
Meanwhile, traditional media outlets like newspapers, books, and TV are grappling with the issue of obtaining payment for the use of their copyrighted material to train AI models, such as ChatGPT and Bard. The distinction between training a large language model (LLM) and infringing on copyright protection is crucial. LLMs deduce and infer patterns in language by analyzing vast quantities of text created by people, but they do not retain the exact content. It raises questions about intellectual property rights and ownership when AI-generated art becomes a reality in the future.
While the evidence for generative AI creating art is currently limited, it is only a matter of time before AI produces artistic works. This poses an intriguing perspective regarding the ownership and copyrights of such AI-generated art. The issue carries implications that will likely lead to substantial debates in the near future.
Addressing the concerns raised by Bender, Hanna, and industry professionals is vital. AI-generated discrimination, deep-fakes, wage theft, and the potential infringement on intellectual property rights demand attention. Moreover, the impact on various industries, including the entertainment sector, requires careful consideration. As AI continues to advance, striking a balance between progress and ethical responsibilities becomes imperative.
In conclusion, the real threats associated with AI focus less on the hypothetical dangers to humanity and more on the existing consequences affecting society today. Proper regulation and safeguards are necessary to address the issues of discrimination, deep-fakes, wage theft, and intellectual property rights. The intersection of AI and human life calls for a nuanced and comprehensive approach to harness the benefits while mitigating the risks.
