What might the rise of generative AI mean for the internet?
The risks of Artificial Intelligence (AI) has primarily been discussed in terms of apocalyptic sci-fi scenarios where robots accidentally kill off humanity, or where lethal autonomous weapons go rouge à la the Matrix, or whatever other way tech/venture capital “thought leaders” like Elon Musk think AI can doom the world. The more measured, and insightful, thinkers of AI tend to consider the risks and harms associated with what AI can do now, rather than what it can hypothetically do in the future. What does the rise of artificial generative intelligence mean for the information super highway? Is it doomed to fall to the charlatans and trolls or can we forge a more hopeful destiny for the internet?
AI is determined by the data it is trained on which means that it can often ‘inherit’ bias in the underlying data. Consider the scandal when one of Google’s AI image generating models, Gemini, was adjusted in order to be more inclusive. Many AI models typically portray high income professions like lawyers and doctors as white men and low income professions like cleaners as black and Asian women due to socioeconomic and historical factors in the societies where the training data came from, and Gemini had inherited that bias. Google tried to fix that by making the model more likely to portray non-white people, which became an issue when people asked for an image of a 1943 German soldier and got a picture of a black man. As I often repeat on the blog, AI doesn’t actually think and can’t actually learn why so it cannot understand why portraying diversity is good in some contexts and distortions of the past in other contexts.
That fundamental inability to truly understand greatly limit the positive everyday use case of AI, but it doesn’t necessarily make a difference for users interested in using AI in bad faith. Indeed, inaccuracies about history, politics, and current event can be a great help if your goal is to troll. And let’s face it, that it the prime use for AI in its currently limited form. Just like your kitchen scales are not accurate enough for serious chemistry, your 5th grade history book isn’t accurate enough for serious history, current AI writing is only at the level of a casual layperson. That makes it unsuited to professional work and not coincidentally perfect for unprofessional work.
The biggest risk of AI to me is how it can spew out low effort garbage on an industrial scale, and how it can be used to fake videos of public figures. Riling people up online is a time-honoured tradition, as is its nastier cousin: scamming people online. Most scams and bad faith arguments work on scale. The more people see my bad faith Reddit comment the higher the chances that someone will miss the joke and go ballistic; the more people who receive my Nigerian prince email the higher the chances that someone will miss the red flags and wire me money. Like the Eurojackpot the chance of a specific individual falling for your trolling or your scam is low, but if you get enough chances at it you’ll eventually win, but the Eurojackpot is €2 per entry while online scams are practically free to enter.
That issue is compounded by the fact that almost all big tech companies, whether in hardware or software, social media or HR platforms, seem to love the idea of AI and want to be seen as one of the cool innovative firms who are developing their own AIs. They aren’t sceptical enough to consider that AI might be a net negative for them, and that is most acutely an issue for social media firms. After all, if the town square is filled with pickpockets, scammers, and impersonators you’ll move your conversations to a more closed off forum even if it limits your reach. And AI is just the latest blow to the 2010’s style big public discussion platform, the pseudonymity and fairly lax moderation policies which are the norm on current social media already empowers bad faith users.
Some people might then ask if the internet as we know it is dead given the current AI-generated garbage onslaught. But I think that is first-level thinking, (more on that in last week’s post) in reality people will not just accept that the internet is dying but instead adapt to the changes and continue to use it, albeit in new ways. To me, the most logical outcome of AI flooding the internet with low effort garbage is that people retreat to private spaces like chats and groups instead of comment sections and feeds. The main beneficiaries will be the people who get more interactions with people they know instead of interactions with anonymous user names. The global free for all that social media has enabled has shown some of humanity’s worst sides. It is too easy to be mean when you are interacting with dozens or hundreds of profiles rather than people. It is too easy to forget that there are thinking, feeling persons on the other side and not internet gnomes who will leave nice comments out for you if you are nice and sour your feed if you displease them.
One of the most famous descriptions of the internet is that it is an information super highway, and the big tech firms spend most of the 2010’s making life difficult for the proverbial mom and pop stores on the byways so that you would stick to a few well-known truck stops. The blog declined in favour of the Substack, the company website declined in favour of the corporate Facebook, and even the torrent sites declined in favour of streaming. Much of that era of streamlining took away power from the people and centralised it in a few tech giants. But the glimmer of possibility exists in this dénouement of AI generation. Maybe people retreating away from the Facebook and TikTok feeds into private chats will mark a resurgence of broad-casting journalism, and maybe it will mark the end of the content recommendation algorithm’s iron fisted rule. Few terms feel as intrinsically optimistic as ‘the information super highway’, maybe it doesn’t matter that there are bandits on the byways as long as the highway can keep us truly connected, as in one person to another rather than keeping everyone connected with everyone else. One can hope, I certainly do.
If you liked this post you can read a previous post about student protests here or the rest of my writings here. It'd mean a lot to me if you recommended the blog to a friend or coworker. Come back next Monday for a new post!
I've always been interested in politics, economics, and the interplay between. The blog is a place for me to explore different ideas and concepts relating to economics or politics, be that national or international. The goal for the blog is to make you think; to provide new perspectives.
Written by Karl Johansson
Sources: https://www.theverge.com/2024/2/21/24079371/google-ai-gemini-generative-inaccurate-historical
Cover photo by Taras Makarenko from Pexels, edited by Karl Johansson
תגובות