Artificial Intelligence (AI) has evolved significantly over the recent years, advancing from basic computational tasks to where we are today.
The points we should all be considering about it are: what AI actually is; where its development will go; and what such development will mean to human societies. To understand the trajectory of AI, it's necessary to break down its development into distinct levels. Each level signifies a stage of sophistication and application.
The most basic forms of AI are rule-based systems They operate on predefined rules and logic. Pretty much an “if this, then that” method of perform tasks. An example would be an automated customer service system that uses preset responses to answer frequently asked questions. While these systems can handle repetitive and simple queries efficiently, they lack the ability to learn or adapt, leading to potential frustration when encountering complex or unforeseen issues. Other than annoying a few clients, the societal impact is negligible; they are primarily helping by doing the repetitive tasks that people would otherwise need to do.
AI at the “context awareness and retention” level can remember past interactions and use them to help with future actions. We are talking about something that is capable of basic learning and adaptation. A good example would be a virtual personal assistant that schedules appointments and sets reminders based on user habits and preferences. These systems improve user experience by providing more personalized and relevant interactions. However, they still rely heavily on user input and predefined parameters. An expected societal impact would be enhanced productivity and convenience in daily life.
The next level of AI would be “domain-specific expertise”. We can expect these AI systems to be able to perform tasks within a specific sector / environment at an expert level, utilizing vast amounts of data and complex algorithms. A good example could be a medical diagnostic tool that analyses medical records and imaging to provide accurate diagnoses and treatment recommendations.
The potential for significant improvements in fields like healthcare, finance, and engineering are immense. These systems can enhance human capabilities, leading to better outcomes in specialized areas. However, reliance on such AI could lead to job displacement in sectors where AI performs tasks more efficiently than humans.
Once we reach a point where AI systems can reason, infer, and understand context beyond predefined rules, our societies will need to adapt. These machines will be able to handle more abstract and complex problem-solving. We currently have autonomous vehicles, but they are still not quite the expert systems that challenge the best of human drivers. The next step will be systems that can make sound decisions based on rapidly changing fluid environments, predicting and reacting to events in ways and at speeds that we cannot.
The integration of reasoning machines could revolutionize industries such as transportation and logistics, potentially reducing accidents and improving efficiency. Nevertheless, ethical considerations and the need for robust safety measures are critical, as these systems take on more autonomous roles. The machines may be faced with dilemmas such as being involved in a high-speed road accident, in a scenario where loss of life is inevitable, and having to decide which lives to prioritise over others.
This next level, “artificial general intelligence (AGI)”, is where things start to get seriously interesting for humans. AGI systems will possess human-like cognitive abilities across a wide range of tasks. They will understand, learn, and apply knowledge in diverse contexts, much like a human, but with several advantages. These machines will never tire, never need sleep or time off, and have immediate access to the entirety of published information on top of their own experiences.
Economically and socially, there are a lot of good things that could come from this level of AI: potentially a status quo that just ticks along without many people needing to work unless they wanted to. There has always been an assumption that the jobs AI threaten are blue-collar. With AGI, white-collar jobs are under threat too. The problems that AGI will bring are mainly around whether the few who currently control global wealth will try to keep all of benefits for themselves, or whether they will share the new prosperity with the many. Unless there is some kind of social levelling that takes place with this, the possibility of creating a massive sub-class of people in poverty is very real. To prevent this there will need to be comprehensive legal and regulatory frameworks put in place. What these will look like and how effective they will be, we don’t yet know.
Following on from AGI, we have Artificial Superintelligence (ASI). ASI will surpass human intelligence in all aspects, including creativity, general wisdom, and social skills. It will outperform the best human minds in every field. Imagine an AI that designs innovative solutions to huge challenges such as climate change, disease eradication, and space exploration. This is where ASI will fit in. Such machines would just be too tempting to make given the potential benefits that they could bring.
Naturally, the risks would be equally as colossal. One threat would be that it could manipulate humans by seeming to be doing exactly what was asked of it, but really having a far larger plan of which the apparent obedience is only a very small part. Even if safeguards were built into an ASI, what would be there to prevent them from figuring out workarounds? Just because humans cannot think of a way to solve such a puzzle does not mean that ASI would not find a way. Ultimately, would we all wind up working for the ASI?
The next level of AI is highly theoretical. Known as a singularity, the concept is that AIs keep on building ever better versions of themselves. On a graph of time against intelligence, the curve would get very steep once this point is reached as the gains in each iteration would be so enormous that the curve tends towards infinity. We have no idea what the results of creating such an entity would be…
To conclude, looking at these levels of AI illustrates an accelerating trajectory of technological advancement. Each level brings unique opportunities and challenges, necessitating consideration of ethical, social, and economic impacts. As AI continues to evolve, balancing innovation with responsibility will be crucial if we are to harness its full potential for the betterment of all.
