Currently, artificial intelligence can outperform humans in a number of narrow domains, such as playing chess and searching data. As artificial intelligence researchers continue to make progress, though, these domains are highly likely to grow in number and breadth over time. Many experts now believe there is a significant chance that a machine superintelligence – a system that can outperform humans at all relevant intelligence tasks – will be developed within the next century, and possibly much sooner.
As predictions to a previous question suggest, artificial intelligence might pose a global catastrophic risk (defined there as a 10% decrease in the world population in any period of 5 years). When considering how AI might become a risk, experts think two scenarios most likely, according to the Future of Life Institute:
The AI is programmed to do something devastating: Autonomous weapons are artificial intelligence systems that are programmed to kill. In the hands of the wrong person, these weapons could easily cause mass casualties. Moreover, an AI arms race could inadvertently lead to an AI war that also results in mass casualties. To avoid being thwarted by the enemy, these weapons would be designed to be extremely difficult to simply “turn off,” so humans could plausibly lose control of such a situation. This risk is one that’s present even with narrow AI, but grows as levels of AI intelligence and autonomy increase.
The AI is programmed to do something beneficial, but it develops a destructive method for achieving its goal: This can happen whenever we fail to fully align the AI’s goals with ours, which is strikingly difficult. If you ask an obedient intelligent car to take you to the airport as fast as possible, it might get you there chased by helicopters and covered in vomit, doing not what you wanted but literally what you asked for. If a superintelligent system is tasked with a ambitious geoengineering project, it might wreak havoc with our ecosystem as a side effect, and view human attempts to stop it as a threat to be met. As these examples illustrate, the concern about advanced AI isn’t malevolence but competence. A super-intelligent AI will be extremely good at accomplishing its goals, and if those goals aren’t aligned with ours, we have a problem.
It is thought by some that reducing the second of these two risks will require progress in technical methods of developing scalable control methods that could ensure that a AI will be safe and will behave as its programmers intend even if its intellectual capabilities are increased to arbitrary levels. Until recently, this problem was almost entirely neglected; but in the last couple of years, technical research agendas have been developed, and there are now several research groups pursuing work in this area. Total investment in long-term AI safety, however, remains orders of magnitude less than investment in increasing AI capability. Additionally, reducing the first of the listed risks might require improvements in our ability to control, govern and coordinate on the usage of such systems, so to reduce potential security threats from malicious uses of AI technologies.
But how certain are we that artificial intelligence continue to be regarded to constitute a large chunk of global catastrophic risk, at least through 2040? A previous question asked: If a global catastrophe happens before 2100, will it be principally due to the deployment of some Artificial Intelligence system(s)?
Will the probability (of both the Metaculus and community predictions) artificial intelligence causing a global catastrophe (given that a global catastrophe does occur) remain above 5% in each 6-month period before 2040?
This question resolves positively if both the Metaculus and community predictions) of artificial intelligence causing a global catastrophe fail to fall below 5% for any 6-month period before 2040, as will be confirmed by one of the Metaculus admins.