Yesterday we were treated to a fascinating and enriching meeting led by Dr. Haim Shaul — an expert in applied cryptography for privacy in AI models, and a winner of IBM awards for outstanding achievement in patents and professional papers.
In plain, clear language, Dr. Shaul explained how the developers of AI set out to imitate the workings of the human brain, across five topics: our brain versus the artificial brain; how AI "sees" and learns to understand what is really in an image; the "school" in which the system guesses, errs, corrects itself and grows smarter; how it can be fooled by changing a single small point in an image; and how new systems learn to create on their own — to write, to answer and to invent stories — out of a high-level grasp of the context of words.