The Internet was designed by humans for humans…or so we thought. To the extent that we have used the Internet in our daily lives, there has always been the expectation that what we saw was human-generated. That is, the information, the articles, the comments, etc. were all written by real people for other people to consume. Yet, we are moving to an age where computer-generated content is becoming a larger and larger percentage of what people see online.
Already, there are fake social media accounts which artificially inflate the popularity of some users on these sites. On Instagram, for example, people buy these bots to increase their follower count and the number of likes the receive on their posts. Similarly, we see computer-generated likes on Reddit and Youtube. It is this latter case which can be worrying. To a large degree, it undermines the legitimacy of the approval and disapproval system, since there is no longer the concept of one user-one vote. At least in my experience, there is an inherent trust that comments with hundreds/thousands of likes and up-votes are actually popular. Rarely would I stop to consider that those numbers were fake.
Yet, it would appear that, especially with the rapid development of machine learning and AI, a lot more content on the Internet will be computer-generated. Indeed, some companies actually seem to be embracing this trend. Already, simple news stories are written by programs fed with facts and set to write based off some template. At a recent Quora tech talk which I had the chance to attend, CEO Adam D’Angelo also was receptive to the idea of machine-generated content. He noted that while the technology is still premature, he could easily envision questions on his site being written by AI. Again, he was a bit more hesitant about the idea of computer programs answering questions, since in his opinion, that requires human experiences. But, it still indicates that the online community is positive about the role of AI.
I, however, remain skeptical about the potential dilution of the Internet resulting from a plethora of computers generating and posting content. It’s already hard enough to sift through the pile of content posted by people on social media. Imagine trying to decipher the truth and/or the relevant information when you don’t even know whether it was written by a human…
An interesting set of ideas…
Of course, the Internet was originally designed to allow computers to talk with other computers; it was only after the initial network was running that the traffic started to be dominated by people talking to people. Maybe the current trend is just going back to the original purpose…although having computers generate the content that is then read by people does seem to be a new and different thing.
I do think there are some differences in the various ways computers are generating content. I agree that bots generating “likes” and links to artificially boost the visibility of a post or sight is not what was intended by the sites themselves. There is an on-going war between the “search engine optimizers” and Google engineers, for example, to try to keep the page ranks immune from this sort of manipulation.
This seems somehow different from AI bots writing simple news stories. Given the worry that human beings always introduce bias into a story, such bots might be the way to keep the bias out (or at least consistent). I’m less worried about the bias of reporters than many in the class (although I do worry about distinguishing between “real” reporters and those that just post), but if that is a huge worry maybe bots writing the news is a solution.
Interesting times in which we live…