Law

Artificial Intelligence (AI) A Threat to Humanity (Part 2) – How AI Really Works

I think I’ve sussed how artificial intelligence (AI) really works.  Particularly when it comes to creative tasks, such as writing an essay.

If I ask AI to compose my next Eurovision hit, it will probably get back to me with a mishmash of every winning entry which is ever been written, including words and music from ABBA’s Waterloo.  If I ask AI to design me an iconic London building, it’ll probably get back to me with the plans and specifications for The Shard.  How do I know this?

Because when I recently asked to AI to write me some promotional material for a book which I had published, it repeated back to me my own words.  So what is the issue?

Only that if you use AI to write your university treatise, you will probably be poaching on someone else’s copyright material.  It’s like paying someone else to take your driving test.  You might even be sued by the copyright holder.  And think about this.  Isn’t it possible that copyright in any text or artwork produced by your AI belongs to the AI machine which generated it?  You might even be sued by your own computer.  But that’s not to say that creative AI is without its uses.

If you are intending to sit a professional examination, you might want to ask your AI what questions are likely to come up, based on past form.  It might even know what questions are going to come up, if it already has access to your examiner’s password protected materials. Not just the questions. Maybe also the examiner’s model answers. But can you trust it to give you the correct information?

Remember the American attorney who used AI to write a skeleton argument which he later presented to the court in a big commercial law-suit.  When the opposing lawyers looked into it, they discovered that every case which the attorney. had cited in his skeleton argument was fake.  The AI generated case law which he had quoted to the court had never even existed.  How embarrassing is that?