Best wishes for a wonderful 2018!!
- Last year (there, I’ve said it!) I mentioned that the “black box” nature of AI predictions poses several problems, not least of which is the need for courts to understand a decision process in order to assign liability when things go awry. Late in the year, several efforts were underway to have AI itself try to solve the problem, and some looked quite promising. The author of this list of “10 Alarming Predictions for Deep Learning in 2018” (see #6) argues the problem may be intractable, and expects “fake explanations” to appear. (Also, #10 is a quick summary of where we stand with ethics in AI.)
- In 2017 Elon Musk repeatedly warned about the “existential threat” of AI. He is much more sanguine about using AI-related tech to supplement the brain via neural implants. “(Musk) wants to use (brain-computer interfaces) BCIs in a bi-directional capacity, so that plugging in could make us smarter, improve our memory, help with decision-making and eventually provide an extension of the human mind.” “He wants to directly tap into the brain to read out thoughts, effectively bypassing low-bandwidth mechanisms such as speaking or texting to convey the thoughts.” DARPA is seriously involved in this technology.
- Here are four very high level tips for marketers re AI in 2018. (As usual, it’s all about the data.)