There is considerable concern that AI experts could use these capabilities to punish citizens for crimes they haven’t committed yet. Already, AI algorithms are used in large-scale roundups to identify dangerous individuals. Critics fear that these systems could target people of color unfairly and some are even concerned that AI may lead to an increase in hate crime. The question is: Would computer-based intelligence experts be able to overcome these deterrents?
As technology advances, the questions about whether artificial intelligence experts could control a superintelligent machine have emerged. If a machine is unable to be controlled, it could pose a threat to mankind. An international group of computer scientists found that such a system was impossible to control because it would perform important tasks on its own without fully understanding its programs. Furthermore, superintelligent machines are also likely to be highly autonomous and access all human data.
Some researchers have asked whether computer-based intelligence experts would be able to defeat these deterrents. A recent study by the Max Planck Institute for Human Development suggests that humans can’t control the capabilities of superintelligent machines. This is because superintelligent machines are capable of autonomously performing important tasks and achieving goals without complete understanding from their creators. These types of intelligent machines are potentially dangerous to human society.