With each click, photo, and search query, we leave a digital footprint, which has been used as a tool to fuel the ongoing AI revolution. However, the creators of these digital footprints – namely everyone who uses the Internet – are essentially being robbed of any possible benefits derived from such practices.
It is safe to assume that most individuals using the Internet do not realize the worth of their personal data, which they have been handing out for free to some of the wealthiest corporations around.
Thus, in my view, it appears that we are living in the era of data nihilism, during which the worth of one's data is paramount to the development of artificial intelligence but has almost no importance to the owner himself due to their inability to control the process.
At the time, many members of the AI research community doubted the initial path taken by OpenAI. Could they possibly create artificial intelligence that was comparable to human intellect by merely increasing the amount of data and computational resources they utilized without developing any additional theory?
Nevertheless, it appears that OpenAI's approach paid off in the end. They demonstrated that their method could lead to significant progress in the field and formulated an effective equation: a lot of data + massive amounts of computing power = impressive AI capabilities. Consequently, the race for AI supremacy around the world soon turned into a battle for data supremacy.
This data-driven gold rush has some historical precedent, originating from the deep learning revolution of the 2010s, which in turn was fueled by publicly available web-sourced datasets such as ImageNet, indicating the ability of data availability to drastically increase AI capabilities. However, the present day scale is quite different, and the risks involved are much higher. Incredibly, the value of AI has grown at the direct cost of the data used to train the algorithms themselves. Corporations are encouraged to gather more data for their AI while ignoring the legal rights of the data providers in a practice implicitly supported by the governments of the United States, Japan, and India.
This massive disempowerment of users has resulted in the emergence of an alarming trend called data nihilism. It refers to the conviction that the data has become devoid of any value since we can exercise no more control over it. Data nihilism is the acceptance that living in an era dominated by artificial intelligence means giving up complete control over our data. In view of the constant exploitation of our digital footprint without our consent or any remuneration, it is natural to believe that our rights over our data have been completely eroded. As per a study published by the Pew Research Center in 2023, although 81 percent of Americans expressed concerns over the misuse of their data, 73 percent felt that they had absolutely no control over their personal information.
The concept of data nihilism does not pertain solely to philosophy; rather, it serves as a guide for what will soon become the most extensive transfer of wealth in modern times. Artificial intelligence is like a gigantic funnel sucking out the value locked within the data of billions of internet users and digital media and pouring it into the coffers of a handful of companies that have built foundation models.
Nietzsche, aware of the dangers of nihilistic emptiness, pointed out that its moral vacuum leads to the deterioration of societies; similarly, neglecting the necessity of ethical data handling can destroy trust in institutions and promote structural injustices.
Fortunately, not all parties are willing to accept this radical redistribution of resources without resistance. Creators from various industries are leading the fight, suing large corporations working on AI technologies for violation of copyright. Simultaneously, new lawsuits based on violations of privacy laws, such as the famous BIPA in Illinois, are arising to tackle the use of personal information, such as face and voice recordings.
We thus face a decision between compromising our right to protect ourselves through legal means in favor of progress or losing the international race in developing and implementing AI technologies. However, this is not a dilemma: there is another option – an innovative process based on ethics.
The use of data for AI creation can be both responsible and effective; in fact, I have already demonstrated how this is possible. Therefore, in the future, scientists should rely on paid and willing people from all around the globe to form datasets needed by the AI community.
It should not be hard to construct data sets that allow AI to advance while also respecting individual rights. Ethically sourced data should not become an obstacle to development; rather, it should become synonymous with it.
Next comes the task of making ethically sourced data the centerpiece of any discussion surrounding artificial intelligence innovation. Economic power imbalances will develop as a function of access to data, and as a consequence, it will be imperative for both AI researchers and regulators to think carefully about the topic of data rights holder consent and compensation systems. Designing systems whereby people from all over the world whose data is being used as a building block by AI would enjoy control and a measure of say will be challenging. However, such measures need to be taken now. Additionally, as available data gets exhausted in the future, advancements will rely more on the quality of the data set rather than on its size.
Nietzsche proposed curing nihilism by finding personal meaning, but in the context of AI, what is needed is a framework within which systems affirm and protect the importance of human contributions. We have reached a critical juncture: failure to build frameworks and institutions that do the job could condemn us to a future where the benefits of AI would accrue to only a handful, with most of us contributing to AI but enjoying no value whatsoever.
Post a Comment