What should we think of it now? Of Artificial Intelligence? As technologists, we know that many AI messages aren’t about AI, but about smart routines and algorithms. It’s been a technological development for 30 years. You don’t often encounter real AI agents, according to Gartner. And Large Language Models (LLMs) are actually based on only 0.01 percent of all available information in our world when they formulate their words. Word by calculated word. We take hallucinations for granted – if we recognise them. Interacting with such a model often feels pleasant and provides you with quick answers.
A deeper dive, beyond funny videos of yourself superimposed on a historical figure, talking to a Yeti, or walking in the dystopian world after the First Nuclear War. We’re now well acquainted with the marketing blunder of WeTransfer – free transfers in exchange for your file and intellectual property to train their AI processes. But what do you do when recruiter Zara approaches you online? You can hire her, she’ll conduct interviews and find you a suitable job – apparently with a much higher success rate than regular human versions of her (m/f). See the human recommendations based on her hard work!
How do you react when the company ClearviewAI also uses your face to train its AI routines? Or when private company Palantir collects your data for the government to analyse for its intelligence service? It’s war, Dutch reporter Huub Modderkolk already wrote, and this digital war has reached many a private PC belonging to the average citizen. Kees Verhoeven – chair of the KNVI Advisory Board – calls these companies ‘digital wolves in sheep’s clothing’.
Users are sometimes thrown off WhatsApp, (like Dutch journalist Ruben Altena), their data policy is questionable, and all sorts of things are being fed into their AI training routines. And then there’s Meredith Whittaker, (President of the Signal Foundation). Signal is suddenly starting to gain traction as a messaging app. They handle the privacy of messages in a completely different way. A different business model. During the coronavirus pandemic, people who questioned government actions were immediately using this channel. Now many others are joining in.
We can already do quite a lot with AI-like and true Gen-AI tools. We might be able to solve labour shortages. We can rapidly develop new materials that help us in construction and logistics. The balance could tip in favour of AI for good. When people act with integrity, they create AI with integrity. Perhaps based on universal principles. Asimov had already come up with a few. Of course, that doesn’t yield a cent for those in power with a thirst for war, mining and steel construction, and the arms industry. But who knows, perhaps with our profession, we can continue to help humanity on the right path.
Smart Humanity – when technology is used to support the Human race (and not the other way around!).
Wouter Bronsgeest
(President KNVI)