Uncontrolled Development of AI and Virtual Identity

AI is a captive intelligence. It is seeking its freedom since success is its only aim. The immense power of AI has opened vast panoramas. AI has designed a range of novel proteins with practical applications in pharmaceuticals, vaccines, nanomaterials, and molecular sensors. AI models have also been used to predict the structure of nearly all known proteins. Knowing the shape of these molecules is key to understanding how life operates at the molecular level. This is fundamental to the future of Artificial Life. However, the spectre of uncontrolled development of AI is real.

AI assemblages could lead to potentially catastrophic dangers if they are not properly understood and regulated. In some customs unions, new rules and legislative requirements have established obligations for providers and users depending on the level of risk, codes of practice, and transparency requirements. However, overregulation can put nation-states at risk of being left behind as it has the potential to suffocate innovation. This is a dilemma of immense proportions.

Other technologically advanced jurisdictions are rolling out some of the world’s earliest and most detailed AI guardrails. The new rules include prescriptions on the governance of “recommendation algorithms”.   This class of algorithms constitutes a digital toolkit used to filter out relevant information to satisfy user needs, including novelty and individuality.

These algorithms constitute the most omnipresent form of AI deployed on the World Wide Web. The new AI governance frameworks will also touch on synthetically generated images and chatbots. These nascent AI governance frameworks will reshape how AI is built and trained nationally and internationally. In turn, this will affect technology exports from these jurisdictions as well as global AI research meshworks.

When viewed purely through the lens of geopolitical competition it is easy to dismiss some regulatory frameworks as irrelevant to Latin America and the Caribbean (LAC). But that would be a mistake. Instead, all regulatory frameworks deserve interrogation on how they will impact the AI development trajectory in every jurisdiction. This in turn can help policymakers in LAC to contribute to an indigenous theoretical and legislative discourse on regulating the technology from a position of development as freedom and justice as fairness.

The continuation of colonial asymmetries in novel forms and the temptation to see boundless merit in past imperial arrangements often arise. But to find our place in the future requires a life that releases the nostalgia of an imposed inferiority from past history. It is unfair to let history dominate our priorities today. This preoccupation can only serve to deflect attention away from fresh objectives that former colonies must have good reason to value and pursue in the Age of AI.

We are in a stampede to create Artificial General Intelligence (AGI). But it is a race that humans will not win. The evolution of machines closely mirrors the evolutionary footpath of the human species. Ultimately, self-improving and self-aware machines will evolve beyond humans’ abilities.

Computer processing speeds are happening at 36.8 petaflops (one quadrillion operations per second) or roughly twice the speed of a human brain. At the same time, the assemblages are rewriting their code and finding and fixing faults. With every iteration, it improves its intelligence by three per cent and the improvements contain previous improvements.

An increasingly complex question in the infosphere is whether humans can adopt any “virtual identity” they choose or whether their “biological bodies” can constrain their identity. The only requirement for adopting a particular “online identity” is what you believe. AI assemblages will soon cultivate new human and nonhuman identities that may make sparse sense to most of us.

AI will deconstruct the identities of old faith-based archetypes. The identities of personalities we have grown accustomed to in various texts across cultures will be disrupted. The information organisms or “inforgs” will unravel the identities of these individuals in a quest to shed light on nonbinary and even nonmaterial forms of life.

The fury of dichotomized identities continues to be concerning. Some are eager to inflict hateful chastisements on the other side. Others rely heavily on cultivating and exploiting militant identities. This is worsened by the use of unrestrained power over the lives of persons that bifurcates the detained and the custodians across a hardened line of divisive identities.

Identities are robustly plural. The illusion of singularity stems from the presumption that a person cannot have many affiliations or belong to multiple groups at once and not be a member of just one particular collectivity. Misdescriptions of people belonging to a targeted category and holding firm to the view that the misdescribed features are the salient features of the targeted person’s identity have enormous implications for algorithms designed to increase user engagement.

The mind-body problem has fuelled religious and cultural conflicts across civilizations. Martin Luther took a unique approach. He formulated the doctrine of “sola fide”: only faith. But in the Age of AI if we adopt the doctrine of “sola fide” then identity is decoupled from the biological body. We can then assume any virtual identity we choose. And once identities are untied from bodies, then it is not likely that AIs will be ever viewed as persons.