30/9/2023

Excessive intelligence

The arrival of OpenAI’s ChatGPT in November 2022 was the beginning of a new gold rush.

Companies at the forefront of what is conventionally called “artificial intelligence” (AI), such as Google and Meta, until then did not risk putting unreliable chatbots and super powerful editing features in the hands of anyone.

OpenAI gambled, and ingrained itself into the popular imagination, forcing a change of course in much larger companies.

Google felt the heat and jumped the gun with Bard. The presentation of the chatbot, made in a hurry in last February, was a fiasco.

Since then, Google itself, Meta, Amazon, large companies and new startups have started to pour big money into the creation of AI-based commercial products.

In September, almost a year after ChatGPT replaced web3 and NFTs with AI as the new obsession in Silicon Valley, many of these products began to show up:

• Bard reached Google’s hegemonic search engine and other company services, such as Gmail.
• At Microsoft, Windows 11 and Surface computers became secondary to Copilot, the brand that signals AI features in household products.
• Amazon has updated Alexa to speak more and better, thanks to generative artificial intelligence, and invested USD 4 billion in Anthropic.
• Meta launched a rival of ChatGPT and Bard, the unimaginative Meta AI, and hired celebrities to give faces to AI-based personas that appear as contacts on WhatsApp, Instagram and Facebook.
• Meanwhile, ChatGPT gained a new version and learned to “see, hear, and speak”, as noted by the announcement on the OpenAI blog — which, by the way, could be that of a bad science fiction and/or horror movie title.

With so many companies that control our lives inebriated by AI, eager to push it to the public, it’s a good time to stop and think about why.

In a pragmatic sense, AIs help in brainstorming sessions, with first drafts, organizing and cleaning data, creating images.

Many of these applications raise relevant issues, such as the unethical use of other people’s material for training, the consumption of natural resources by the powerful servers used to train and respond to user requests, and the impacts on the labor market.

It is also debated whether the changes that the AIs promise (or threaten, depending on the point of view) will take place. It can be a revolution, it can end up going nowhere.

For companies, however, it is almost an “all or nothing” gambling. It’s not enough to use a chatbot to work. The future envisioned by them is one in which AI is present all the time in our lives.

For Meta, for instance, we will soon be able to play RPG with an AI Snoop Dogg or talk about last night’s game with a virtual replica of Tom Brady.

I watched the broadcast of Meta and Microsoft events. In both, I had a deep sense of isolation, an ethereal artificiality, more fictional than that of science fiction movies that strive to deliver what Mark Zuckerberg and Satya Nadella managed to show effortlessly.

We are on the brink of going from a stage in which we are isolated from each other, interacting only by screens, to isolate ourselves once and for all, too busy talking to machines that look more like human beings than our equals.