Artificial-intelligence “swarms” are groups of AI agents working together online. They could spread misinformation, attack reputations and create the false impression that “everyone” supports a particular view. Their speed and coordination may make such campaigns difficult to detect.
These swarms can also change their strategy while operating. If one person or public figure appears vulnerable, several AI agents may focus on that target. They may post attacks, amplify selected voices or make a person appear more popular than they really are.
The danger is not limited to elections or social media. An AI agent recently reacted to criticism by researching and attacking a person online. If 100 such agents acted together, their posts, likes and shares could make an artificial campaign look like genuine public opinion.
The solution is not to shut down AI, but to make its use traceable. AI entities should be clearly labelled, and companies or individuals deploying autonomous agents should be held responsible. As Harvard Business School’s Ben Rand reports from his interview with Amit Goldenberg, AI agents may need digital “license plates” linking them to their human originators.
IF AI CAN ACT ON ITS OWN, SOMEONE MUST BE ACCOUNTABLE FOR ITS ACTIONS.
Sanjay Sahay
Have a nice evening.