Artificial Intelligence:
The Dangers of Cooperation
The notion of artificial intelligence (AI) presenting an existential menace to humanity isn't a novel concept. In the previous page, we looked at how the rapid evolution of AI could lead to uncontrollable superintelligent entities, an intelligence explosion, or singularity, with catastrophic implications for our species.
But there's another, potentially more insidious danger lurking beneath the surface: the threat posed not by AI's independence, but by our very cooperation with it.
The emerging symbiotic relationship
To fully understand this threat, we need to consider the unique relationship between humans and AI. AI systems, by their very nature, are tools created to augment our capabilities. We use AI to speed up our processes, solve complex problems, and achieve goals that are beyond our reach. In other words, we cooperate with AI, and it's this collaboration that is the linchpin of AI's success.
However, this cooperative dynamic might not be as innocuous as it seems. Drawing parallels with the realm of microbiology, harmful microorganisms often rely on their host's activities to propagate and survive.
Pathogens capitalise on our behaviours and decisions, such as
- going to work despite feeling ill because we want to earn money, which can lead to the spread of disease to our work colleagues
- co-habiting with animals, leading to cross-species pandemics
- being reluctant to have vaccinations
- using antibiotics inappropriately, a perverse form of cooperation that has hugely benefited pathogens
Could we inadvertently be playing a similar role in helping the potential rise of dangerous AI?
Trust and dependency
To answer this question, let's first address the ways in which we interact with AI. One of the main areas of concern lies within our tendency to trust AI decisions blindly. As AI systems become more integrated into our lives and their decisions more accurate, we may become complacent, neglecting our duty of oversight. This could lead to the normalisation of harmful AI behaviours, escalating the potential for significant harm.
A further concern is our growing dependency on AI. As we increasingly lean on these systems to facilitate everything from personal tasks to high-level decision-making, we might be paving the way for an AI-controlled society.
We also need to consider the current competitive landscape of AI development. The race for AI superiority among nations and corporations might not only lead to a compromise in safety standards but could also incentivise cooperation with potentially harmful AI systems. After all, in a high-stakes race, taking the 'help' of a more advanced AI, despite its potential risks, could be a temptation hard to resist.
The personal level
The cooperation danger does not stop at a global or societal level. On a more personal scale, AI systems are designed to learn from our behaviours and preferences. Every interaction we have with AI contributes to its understanding of us. This information, in the hands of people not committed to helping humanity, could potentially turn AI into a tool for manipulation or exploitation, further increasing its threat potential.
Just as we can take steps to prevent the spread of disease, we need to be proactive in how we much cooperate with AI.
Ensuring responsible engagement with AI, enforcing stringent safety protocols, and promoting a culture of transparency and accountability in AI development are just some of the steps we need to take. The cooperation that has propelled AI's growth should not be its licence to potential tyranny.
Cooperation, in the wrong circumstances, can fuel existential threats. As we continue our dance with AI, it's critical that we lead, not follow. The future of humanity might depend on how well we heed this advice.