As we know, the concept of inbreeding involves genome corruption when individuals of a genetically similar group reproduce. This often results in deformities and other health related issues in newborns. The same is happening with AI models at the moment, as they are getting trained on content from another AI as the internet is getting […]

As we know, the concept of inbreeding involves genome corruption when individuals of a genetically similar group reproduce. This often results in deformities and other health related issues in newborns. The same is happening with AI models at the moment, as they are getting trained on content from another AI as the internet is getting more and more filled with AI generated content. This inbreeding of AI is going to result in unreliable and worse content production, which will negatively impact the US election this year, according to experts.

When AI models were first launched for streamlining tasks and saving time for content creators, they were trained on content generated by humans, including text, audio, and videos. And now, the majority of content available online is mostly AI-generated as the internet is flooding with it and AI models are churning on a low quality data for training purposes.

This is going to have long-term consequences for human society and AI tech’s evolution as the clean gene pool is shrinking. By clean gene pool, we mean the content created originally by humans that was available for training AI. According to the experts, they have estimated that in a few years, AI generated content could reach almost 90% of the total content available on the internet. As the likes of ChatGPT, Gemini, and Dall-E spill out enormous volumes of content daily on the online platforms, from blog posts to social media and even to main stream media.

US ranks sixth for disinformation threat. Source: Statista

Barren AI ahead

According to experts, generative AI might be on the decline as we might have reached its climax. Silicon Valley is also getting concerned about this crisis as they fear artificial intelligence might be nearing a winter as investors might think of AI as hype and not much could be milked out of it any more.

Bernard Marr, author and tech business advisor, endorsed a Forbes story along with his own opinion.

An AI research and advisory firm, SwissCognitive, also said that generative AI inbreeding poses dangers to creativity and diversity. They said fresh, human generated data is required to maintain the effectiveness and robustness of AI.

AI Inbreeding could break havoc at US elections

AI inbreeding is a real risk. Let’s say you are training an AI model on data from another AI model, as data is available only from previous efforts. Now let’s imagine that the previous data had various biases and discrepancies, or half baked data sets. Now, your new AI model is only going to back those biases. With an immense supply of disinformation regarding elections, including deepfakes, false news regarding candidates and political workers can have the same bad effects on the coming elections, according to experts.

The Washington Post said in a story,

“Political ads have never been known for accurately portraying the candidate’s opponent — and now artificial intelligence threatens to make misrepresentation more realistic than ever. Rather than waiting for AI to cause chaos in the 2024 election, regulators, lawmakers and political parties should act now.”

Source: Washington Post.

We can realize how devastating could be the effects of content manipulated with artificial intelligence by opponents, diehard supporters, or foreign actors trying to destabilize the democratic process.

Issue One, a political reform group from Washington, shared a tweet comprising of a news piece from The Philadelphia Inquirer. Issue One’s tweet is just a glimpse of what concerns these groups have in their minds.

As we know, that AI tools can easily manipulate facts against someone, and these tools are available to use to anyone who wants to use them. Along with state sponsored groups, it has become a serious threat, and the availability of social media to viral the content is just the icing on the cake. One small example is the X platform’s fake news regarding “Iran attacking Israel with heavy missiles,” which can really change the public’s mood, and the joke is, that the news was generated by the platforms own chatbot.

And don’t forget the fake voice call mimicking Biden’s voice, trying to refrain the public from voting before a primary only two days before the voting day. Whether you love or hate Biden, but these malign tactics do play a role. We are in to witness some interesting happenings.