The Internet Doesn't Want To Help You Anymore – Does AI?

Does HAL 9000 really know the best pasta restaurant near you

Mar 3

I.

Let me share an odd internet adventure with you. The other day, I could not recall the term “evidence board”. Naturally, I typed into Google the search term “what are the pin boards with faces called”. In case we’re not on the same page, I’m talking about those evidence boards from procedural crime shows and thrillers. Boards a little like this one:

Me trying to explain to Google what I meant.

Google reacted with great confidence, offering the following top results:

Decorative Open Face Cork Boards - Displays4Sale
Outdoor & Indoor Cork Board Bulletin Boards - Displays4Sale
Corkboard Shapes - Etsy
Square Pin Board Cork Board multiple Sizes - Etsy

Google wasn’t interested in helping my caffeine patch-assisted brain figure out what on Earth “the pin boards with faces” were called. No, Google wanted to help sell me a pin board. Now, we could rephrase the query to exclude the product keyword – in my case, ‘pin boards’ – but that’s what the kids call ‘small dick energy’. Instead, we can just add ‘reddit’ to the end of the search query, and watch Google’s tone shift drastically.

First result:
- In crime TV shows, there are often boards with pictures and ... – Reddit
Fifth result:
- Do real detectives use those yarn boards to solve cases? – Reddit

Did Google know what I wanted? Still not quite. But while my initial search term was rather vague, adding ‘reddit’ at least got Google in the right ballpark. I could then jump to forums of folks asking the same sorts of questions and getting the right answers. I’m far from the first to discover this, but the existence of this trick indicates a larger problem. Search engines are increasingly preoccupied with serving advertisements, rather than with aiding their users.

But what about AI? Let’s ask.

OpenAI’s ChatGPT, when asked the very same question.

ChatGPT, running on GPT-3, manages to give me a term for what I’m looking for and even alludes to law enforcement which, thematically, is in the general realm of my query. There’s just one problem, “face pin boards” is not the correct term. In fact, not only is it not the correct term, it’s not a real term at all. It’s not even a term anyone has ever used, as far as the rest of the internet is concerned. So, Google tried to sell me something, and ChatGPT just lied to me. The internet really is one manipulative character these days.

But it’s about to get worse.

II.

Microsoft, via their often belittled Bing search engine, has announced an integration with this AI technology, named Bing Chat. Google, following reports that they were terrified of ChatGPT, announced their competing product, Bard, just a day after Microsoft. Suddenly, it doesn’t sound far-fetched at all to say that AI chat will largely replace traditional search engines, especially when many smartphone users already perform their searches through digital assistants like Siri. But there’s a problem, and the problem is money. The entire business model of search engines pivots around advertisers bidding on keywords, or as Google calls them ‘adwords’. Chatbots will change the mechanism of search, but they won’t alter the fundamental financial imperatives.

So while search engines jockey for AI dominance, one question remains. How will they continue to monetise internet inquiries? One solution could involve ‘paid bias’. For example, if you asked Bard or Bing where the best pasta restaurant in your suburb was, it would be awfully convenient to the platform provider that it recommended the pasta restaurant paying the most on adwords. After all, Google isn’t a company known these days for neutrality, and their chatbot’s opinions will be qualified as “subjective and potentially inaccurate”. When eight billion Google searches are made per day, however, that subjectivity starts to seem rather influential.

But, as the famous saying goes, this runs deeper than just pasta.

ChatGPT has already demonstrated its lack of political neutrality, choosing which presidents it likes and which demographics of society it is safe to joke about. So what does an internet filtered through the AI hive mind of a mega-conglomerate look like? Consider news media. In the AI chatbot future, we won’t read articles. Chatbots will summarise the story for us in a paragraph, and then we’ll probe with further questions for any extra detail, probably aided by predictive button prompts to save unnecessary typing, or thinking. No need to skim headlines with our thumb, forming baseless opinions off of sheer gut instinct – AI will do that for us.

Dishwashers washed tedious dishes for you, thus saving you the bother of washing them yourself, video recorders watched tedious television for you, thus saving you the bother of looking at it yourself. Electric Monks believed things for you, thus saving you what was becoming an increasingly onerous task, that of believing all the things the world expected you to believe.
Dirk Gently’s Holistic Detective Agency (1987)
Douglas Adams

That’s decapitating to news media as their entire business model is built on clicks. In fact, without the click revenue to justify their editorialising existence, news may come to be fed straight into the machine without any conversion into natural language at all. World news reduced to raw data rendered meaningless without interpretation, nowhere indexable by search engines, written in computer nonsense because it’s ‘more efficient’. News-by-machine might be the dominant source of news in the pretty near future.

That could be a good thing.

III.

Think of it this way. We all are, or at least we try to be, meta-journalists. That is, we journalise the journalists as if we were the journalists and they were our sources. We try parsing the emphasis and omission in their language, the financial stakes and the political bias underpinning their reporting. We aggregate what we can and try to come away with a ‘take’, or – for the kids and Twits – a ‘hot take’. But if chatbots meta-journalise for us, they could be designed to actually give us more dimensions on political issues than we are currently likely to seek, by portraying all sides of the issue.

Sure, we could demand ‘neutrality’ from our chatbots, but how possible truly is that when every word is loaded with bias? Was it a ‘protest’ or a ‘riot’? Well, we might strip subjectivity from the term and simply say it was a ‘gathering’. But now ‘gathering’ is such a squeaky-clean word as to imply purposelessness. The use of unbiased language is, itself, a manifestation of bias, much like omission. Instead, a well-designed chatbot could force us out of our echo chambers and reverse-course some of the toxic pathologies regularly enforced by social media algorithms.

But that’s a little optimistic.

More likely, we will just retreat to our ideological silos. With more and more chatbots popping up, it’s only a matter of time before they go the way of social media and divide along ideological lines. The search engine You is sporting its own conversational search AI, YouChat, which has quickly proven to be less, let’s say, politically timid than ChatGPT. Every political niche will soon have its own brand of chatbot, existing not only to validate its users’ pre-existing views, but to form emotional attachments with them. Even the Chinese communist party will have their gimped rendition. ‘ChatCCP’ will never have heard of Tiananmen Square and will politely let us know if we don’t have enough social credits to open our fridge door.

Trickle-down economics sound pretty good coming from her.

As long as these services are centralised, the question of who controls the parameters will loom over them. The only way for technology to elevate society without building tools for coercion and propaganda is through open-source software and decentralisation. OpenAI started out as a non-profit group, before becoming for-profit. Even now-departed co-founder Elon Musk raised an eyebrow at this curious turn of events.

Centralised or not, however, Reddit and other forums won’t be safe, not once entrepreneurial marketers figure out they can deploy chatty AI’s to astroturf our comments sections. Picture lots of little bots, all chirping their circuits at us – or each other – about the best brand of iPhone case. They’ll spout opinions, invent anecdotes, and share slanderous horror stories of competitor products and rival services. Their uncles will actually know the guy who founded the company and they said that he’s a really good guy who’s really passionate about making good iPhone cases. We’ll be reading the opinions of machines more than those of people, and on that day, the internet really won’t be able to help us.

Because the internet will have died.