Saw this at IDFA 2019, the documentary filmfestival in Amsterdam. If this film wanted to show that we still have a long way to go as far as AI is concerned, it is a great success. To understand human behavior and particularly our language, remains still a problem. We have surmounted a lot of issues in that area in the last 30 to 40 years, but we see here very clearly that proficiently handling dialogs is far away. In the beginning we see a clear misunderstanding about a common word: verb "to like" versus comparison "looks like".
Another striking example is that the robots have apparently learned a lot of standard opening lines. We see these used sometimes at the wrong place at the wrong time. Like Check's companion saying that Chuck seems to be good in bed, a sentence as misplaced as can be.
Also, the dialogs with the elderly women are evenso lacking a real mutual understanding. Maybe, when the respective "couples" get used to each other after some time, and continuously learn from each other, it can well become more productive that what we saw now. The couples are not really used to each other yet, their relationships are relatively fresh.
There was a good example of ethical problems, like curing cancer. A robot could propose to try 1000 cures on 1000 people, and then decide what was the best remedy, this in the context of what is good for mankind as a whole (what do 1000 failures mean when all people in the world will benefit). However, it is not the strategy we prefer, having it in our own (human) way. It abundantly emphasizes that we must choose a good assignment with the proper "fine print" conditions, to let an AI go its own way and proceed on its own devices. And, of course, there are legal issues, nowadays prominent when self-driving cars are concerned, but that is not the only area where such things are relevant.
All in all, several relevant topics are covered in a way digestible to viewers not primarily interested in SciFi. It gives food for thought by showing that overly optimistic news reports cannot be trusted blindly. It won't reverse people who already think that we can solve everything just by letting capable technicians work on it.
Another striking example is that the robots have apparently learned a lot of standard opening lines. We see these used sometimes at the wrong place at the wrong time. Like Check's companion saying that Chuck seems to be good in bed, a sentence as misplaced as can be.
Also, the dialogs with the elderly women are evenso lacking a real mutual understanding. Maybe, when the respective "couples" get used to each other after some time, and continuously learn from each other, it can well become more productive that what we saw now. The couples are not really used to each other yet, their relationships are relatively fresh.
There was a good example of ethical problems, like curing cancer. A robot could propose to try 1000 cures on 1000 people, and then decide what was the best remedy, this in the context of what is good for mankind as a whole (what do 1000 failures mean when all people in the world will benefit). However, it is not the strategy we prefer, having it in our own (human) way. It abundantly emphasizes that we must choose a good assignment with the proper "fine print" conditions, to let an AI go its own way and proceed on its own devices. And, of course, there are legal issues, nowadays prominent when self-driving cars are concerned, but that is not the only area where such things are relevant.
All in all, several relevant topics are covered in a way digestible to viewers not primarily interested in SciFi. It gives food for thought by showing that overly optimistic news reports cannot be trusted blindly. It won't reverse people who already think that we can solve everything just by letting capable technicians work on it.