Generative AI is a fast-moving target and it is difficult to assess its full societal impact just yet. This does not mean, however, that we can’t speculate on who will benefit from this disruptive technology and who will suffer. Will it be current leaders in tech and higher-educated workers, or could it actually benefit the underdogs?
In recent history, many assumed that AI would eat its way up the career ladder; factory robots and self-driving cars would replace low-skill jobs first. While this has proved true to some extent, the current wave of generative AI applications paints a different picture. We can already see how highly educated workers are forced to adapt. These large models will become more powerful and capable and will be trained on sector-specific data. As a result, they are much more likely to take over white-collar work, which is, by definition, all about processing information. In many ways these kinds of tasks are far easier to automate than blue-collar work which takes place in the unruly physical world.
On the company level, we can see similar dynamics. Big tech companies may have laid the foundation for this generation of AI’s (transformers, the underlying innovation of chatGPT, are developed by Google), several ‘startups’ such as OpenAI and even open-source projects like Stable Diffusion are currently stealing the show. Perhaps Big Tech is hesitant to showcase its most powerful applications and leaves the stage to these smaller developer groups that have not as much to lose; both Microsoft and Meta have already experienced how loose-lipped chatbots have damaged their reputation. Yet, the smaller companies’ bots are currently being trained by enthusiastic users, and Big Tech may struggle to keep up with their development.
Is this too much of a naïve take on the future of AI? In the case of Big Tech, integrating these large language models into their ecosystem of services could give them the upper hand once again; they already have the users, the cloud infrastructure, the financial power and successful services to develop even more powerful and usable AIs. Likewise, on the job front, highly educated workers could very well be most apt to use these tools, assess their outputs and create value out of it. As a result, the biggest fish might become even a little bigger after all.
