iHuman: L'intelligence artificielle et nous

Saw this documentary at IDFA 2019, the documentary festival in Amsterdam. This movie did not bring much news for me. This is not to be construed as a complaint. Rather the contrary, as I'll explain later. Much of this was already a topic of growing concern in IT-related journals. Also, lectures held at congresses and seminars presented ample eye-openers in this field, once I became aware of it and began following relevant specialists in the field. In other words, I was not taken by surprise when watching the very many relevant issues passing by. However, that is me, and it is more relevant to make the general public aware, and policy makers in particular.

I know it is not easy to find the right packaging for IT-related contents. I especially know how difficult it is to find the right visuals to support the message on screen. Of course, we now (again) saw the obligatory amount of screen gibberish (mainly program source text), network cabling, flashing lights on appliances, the insides of a server farm, and other boring images seemingly unevitable in this context. Talking heads cannot be avoided either. Nevertheless, I know of no better alternative to present the alarming message. The filmmakers used sort of an all-knowing narrator who guided us throught subsequent stories. I'm not sure that is the best solution, but it may work very well with an uninformed viewer who will intuitively build some trust in this man because of his reputation.

I asked my companion, not working in IT, rather one of the power-user type, about the eye-opening effects for her. She confirmed that this movie worked indeed and could very well serve its purpose to show interested people the many dangers ahead. Of course, those who are not interested at all, cannot be convinced with either talkshow, movie, book or article, so are beyond hope on all counts.

All in all, I suspect that this movie can do a good job of educating people on these very relevant topics, pertinent for everyone and certainly not confined to the world of IT. We cannot leave the decisions to the technicians who work there. We should particularly distrust their management, only interested in short-term profit, and damn the consequences. Politicians do not know yet how important it is for them to step in very soon, rather than wait until the problems become too big to unravel in hindsight. A few high-profile incidents in recent years (Facebook and Cambridge Analytics, among others), may be considered later as a blessing in disguise, by showing the average man/woman how these matters affect their lives. It remained under the hood for too long a time.

Entertaining and engaging documentary. Comes at this topic more from the angle of what could be, and is, the negative impact of data harvesting and analysis etc. The way it is edited ironically suggests a certain bias, as the extremes of AI and algorithms are shown as Brexit, Trump, Bolsanaro etc. But it provides no smoking gun and ignores the fact that the big tech companies and employees who create these algorithms tilt more to the left. So while 3rd parties manipulating the model to incite "right wing" movements is certainly there for debate, it does little to touch on the broader topic of the bias of the algorithm designers themselves. I liked and enjoyed watching it, and recommend it, but it does itself illustrate the difficulty of avoiding bias.

Personally I've seen nothing new on AI although some given insights by people connected to the industry were quite interesting... The presentation of a Swiss as a 'father of AI' on a mountain hill was pretty amusing as it's just another example how millenials like this doc maker play with history :) AI was found as a discipline in the 50s of the last century and thoughts about raised way earlier.. The part on robotics was also limited to the work of the Swiss 'AI father' and a sight on the Sophia 'humanoid' with just some shots on the better examples available in Japan as well as more interesting visions about implementation of those devices in that country. For some the use of graphics to boost the perception that something fearsome is about to happen might have been helpful :) I wonder what viewers at the IDFA screening picked up when they left the theater as many of them grabbed their phones and most likely used AI driven tools to let their digital friends know what they just have seen...

Slow movie with a lot of presumably competent people offering their predictions of the future life with AI. Nicely done and with a lot of ambience. But its all been told before, and there's nothing new or revealing. Becomes repetitive and kind of boring.

All in all the movie was ok, but there were a few things that could have been avoided. On the good side, it was interesting to see some smart people like Michal Kosinski, Jurgen Shmidhuber or Ilya Sutskever talk about the future of AI. I also have to give credit to some of the scenes, structure and visual effects of the documentary. But that's as far as the positive feedback on my side goes. As of 2020, most of the things I heard, however, were a bit outdated. To make matters worse, I couldn't help but feel that the author resorted to fear to create yet another dystopian vision of the future. The viewer is most of the time threatened with ominous music and spooky animations that are meant to evoke fear. It's too bad because it could have been a decent documentary.