interim a.i. : growing artificial intelligence for projects, pt I

A.I. is not a sudden and profound cosmic leap of technology. AI evolves slowly; it entices and lures us every day, as we lurch awkwardly toward a future doing whatever the hell we want.

Eighteen months ago, the CEO of IBM arrived late to join a broadcast panel of global leadership on the topic of developments in artificial intelligence. She was quick and understated in her clarification that IBM and Watson were really about ‘augmented intelligence (AI2)’ and were not immediately focused on the more abstruse concept of artificial intelligence (AI1) as many have generously defined it for more than two decades.

The difference between AI1 and AI2 is a profound distinction. A peculiar thing happens when we surgically remove and substitute the term ‘artificial’ with ‘augmented’ – and a more benevolent and practical application emerges. The notion of ‘artifice’ is complex; it has stratospheric lower-bound limits and a cloudy definition overall. An 'artifice' is an unnatural (whatever that means) construct – and that requires design (whatever that means) and unwilling (whatever that means, too) matter. Augmentation, on the other hand, may be characterized as ‘anything added that works or not’. Something augmented does not necessarily mean it is artificial; it can be a process, a trick, a tool - or it can be organic. Augmentation is normally quite human.

The term ‘augmented’ can be innocuous – even collegial and happily familiar. It can mean supportive reasoning and extra data storage; it can mean occasional guidance or assistance for the two mental functions we, as people, seek most – thinking and memory.

We educate ourselves for two immediately existential reasons: the first is to get stuff to think over (information and knowledge), and the second is to improve efficiency, capacity and accuracy for our central processing power (thinking) in relation to how the world appears to us.

Augmented intelligence, or AI2 can potentially arrive in the form of a plastic backpack hardwired to our nervous system – empowering us with outsourced but localized processing power and data. But, if we contextualize it in a wider, more natural and organic sense – it can mean a collection of available technology (books, databases, software applications, instrumentation end-devices, radio, cameras, friends, colleagues, teachers, etc.) we leverage to improve two immediately existential needs – information storage and mental power. Everybody is doing it.

AI2 is not so scary in most circumstances - mostly because it is gradual, relatable and non-disruptive – and has been for a few thousand years.

In 2019, as we look around for the next hi-tech start-up, there could be a considerable amount of time before the invention of artificial systems that are wholly superior to organic neocortex (thinking brains). Ray Kurzweil, co-founder of Singularity University, continues to predict a near-term trajectory converging on a cloud-based, nanotechnology-infested existence where, somehow, we can mentally manage an infinity of ideas, memories and continuous ethereal synaptic networking. 

However, as argued in MIT’s Report on Technology (April 2019), the energy, effort and total cost of innovation is escalating exponentially, and each new successful idea and technological implementation is terrifically expensive and demanding. This phenomenon limits the rate of progress to a less dizzying, ‘not so fast, Bud’, evolution of the robotic kingdom.

In the interim, our untethered social media tends to exaggerate the immediate and near-term implications and use the term ‘artificial intelligence’ with little intelligence. It is not uncommon to hear references to anything more complicated than an MS Excel spreadsheet referred to as artificial intelligence, deep learning, machine learning or complex algorithms. And, others near them rarely argue – in hope that no one drills into what the hell they are really talking about.

Spreadsheets and almost every algorithm at work in the universe are not Artificial Intelligence, in its technical meaning – they are increasingly common instances of augmented intelligence. The minimum threshold of what qualifies as artificial intelligence is lofty and indefinite - and is well beyond that which we give ourselves credit for achieving. AI1 transcends pedestrian thought and linearity, exceeding the scope of almost everything anyone does anywhere on Earth in 2019.

There is only one form of intelligence with the current capacity to conceive of AI1 or AI2, and that is AI0 - Anthropic (human) Intelligence. 

Smart phones are not so smart. People are smart. An average student struggling with calculus is smart. We just need a little help sometimes, even though we may not wish for it. After all, without a good memory, a person can delight in sporadic wisdom - uncluttered by excessive facts - and there is no burden of consistency.

R Harms, Iscient (c) 2019 LitA pg