Astral: Start-up to flood Reddit with AI garbage, er, marketing

AI spam is already flooding the internet. Google is struggling to show helpful content in its search result rankings instead of AI-spammed websites full of repetition and platitudes. Social networks are full of AI images and links to bot farms, absorbing all attention with their insanity. Jesus made of crabs? Huh! It's hard not to look. Now AI agents are to take over marketing too. Astral is the first to attack Reddit.

In a promotional video, the start-up talks about how AI agents could take over marketing 24/7, i.e. around the clock. All you have to do is give the agent a few pieces of information and it can autonomously navigate to Reddit in the browser and move around as if it were a human. By taking screenshots of individual pages, an AI model can evaluate the images and then interact with the page. Other companies such as Anthropic and Google have already demonstrated something similar.

With Astral, however, the AI agent does not take on any task that the user assigns to it, but instead automatically creates posts on Reddit: marketing posts. The AI bot even searches for a suitable subreddit and formulates an advertising text that should sound completely human, says founder Savannah Feder. A so-called human-in-the-loop, i.e. a person who checks and approves the posts, at least does not seem to be necessary.

AI spam and the platform operators

Real people who are on Reddit are probably not particularly happy about this. Almost two years ago, they protested against the platform's plans to charge money for access to the API. This would have made moderators' work more difficult, as some of them use third-party services that are connected via the API.

Of course, it remains to be seen how good Astral's contributions are. If you look around LinkedIn, the posts that appear to be dominated by AI do not bode well. In fact, the platform operators, like Google, are generally aware that automated AI content is problematic. This is why it is forbidden to use Reddit's API for this very purpose. The same applies to platforms such as X and Instagram. Savannah Feder also writes on X that it is much easier and cheaper to use the browser than to use the API.

Incidentally, captchas didn't stop the AI bots either. In principle, however, these do not stop people from using AI to formulate posts and then prepare them for the respective platforms and simply copy them over. So there are still no ways for platform operators to prevent AI agents.

Astral is currently in a closed beta phase, according to the video. However, users can already be placed on a waiting list and more users will be activated every week. Extensions that also allow Instagram and Tiktok to be populated with AI content 24/7 are set to follow. At least Tiktok itself already offers the option for advertisers to use AI to have advertising texts read out by an AI person – in a realistic style.

(emw)