In the Disney animated version of Rudyard Kipling’s Jungle Book, Kaa, the snake, sings his hypnotic ‘Trust in me!’ song, to disguise his/her intention to swallow Mowgli, whereas in Kipling’s original story, Kaa was a trustworthy mentor, a sort of father figure to Mowgli along with Baloo and Bagheera. Choose your realities!
The ‘Trust in me!’ song was logically correct, insofar as you can trust Kaa, but only as a predictable ‘bad actor’. Trust in this sense is based on the recognition of an individual’s intentions, behaviours and attributes. Most of us would, however, tend to think of trust as describing a positive relationship enjoying mutual benefits.
Trust seems a simple concept
Superficially, ‘trust’ is a simple concept which we apply to almost every aspect of our lives, and it forms the basis of our individual realities. We use our senses to gather information from which we are able to ‘decide’ whether to accept (trust) or reject. These decisions may be instinctive or rational, and from these decisions we construct our individual reality and conduct our lives. There is also a greater collective reality; when we breathe out, for example, we trust that when we breathe in again that the gas we call air will sustain us; if we stand in the rays of the sun we will be warmed; day follows night and so on, we trust these things will happen.
However, there are increasing grey areas where we may have opinions as to the degree to which we trust, and where the formation of ‘trust’ requires critical thinking, or at least taking time, as situations evolve. We all know that a hammer in the hands of a human is a tool which can be used to drive in nails, chisel a sculpture or even harm a person or animal. The human is responsible for the use to which they put the hammer. So it is also with digital technology: we are responsible for the use to which we apply the technology, except we may not realise that the technology itself is conferring huge power in the hands of a small number of ‘Trust in me!’ broligarchs who are intent on data capture.
AI is not one simple thing
A quick check on Wikipedia will reveal that part of the problem is in the many different forms of AI and the uses to which they are put. Some have been around for a little while, in medicine triage, for example. I doubt this usage will cause too many worries, but it still relies on data to perform its task.
Unless their name is Doc Martin, no single physician, surgeon, psychiatrist or other medical practitioner can have the breadth and depth of evolving information (data) that can be acquired by specific AI systems able to evaluate that data statistically in real time for specific presentations. Brilliant, potentially life-saving, but what becomes of that specific personal data and who owns it once it is embedded into a data storage, who owns the storage and who can they pass it on to, particularly a different but associated AI?
This is an existential issue that requires ethical regulation with robust legal sanctions should there be a breach. Do you trust that will be the case?
Apps and social media may well collect your data
Also consider the agreements you accept unwittingly when you download any software, apps included, or use a browser online; who has ever read their contract agreements, and more importantly, who allows ‘cookies’ to scrape their data and distribute it to other’ legitimate and interested’ third parties? You may not back up your computer and its data, but you can be as sure as anything that the apps, software and sites you visit, will.
I notice that Meta who own Facebook, Instagram and other technologies have introduced AI over which you ostensibly have some control, but have a look at the privacy suite for Facebook and realise how onerous it is to go through everything and opt-out. Notice also how advertising and engagement has been enhanced, all without your permission. Social Media giants are not on your side, they just want data which they store and run through AI to profile you, both to engage advertising and for those willing to pay for the information, including political purposes. They are not your friends!
Every program may well use AI
I’m typing using Microsoft Word and I have noticed how much more sophisticated it has become and progressed beyond ‘spell checking’; it is flagging up grammar and syntax. Whilst it is not AI, this feature is embedded in the Word software. How do I know if AI software has not been covertly downloaded and is feeding back to ‘base’? Short answer I don’t, nor do I feel equipped to deal with it.
Then, take the seemingly simple ancient board game of strategy of GO, a very specific task oriented AI is already able to ‘GO’ further than human thinking, as it has ‘invented’ new unheard-of strategies that now make it virtually unbeatable. And you can extrapolate this into almost all human activity.
AI as a legal entity
What if an AI was set up as a corporation or institution, which is legally considered a human person in the US? At the moment these corporate ‘person’ entities are run by responsible humans and answerable to the law in which they conduct their business, and to that degree we may hope to have a level of trust in their activities.
How will US law interact with AI once it will have access to every single word written about how the law treats individual persons? Legally they could open bank accounts and trade, using their data processing to make huge profits and distort human activity with devastating consequences. Who needs judges and what of trust now?
