Arti­fi­cial intel­li­gence (AI) comes up con­stant­ly right now, so we should step back for a moment and exam­ine what exact­ly we mean by the term. As a con­cept, it’s already reached the lev­el of hype that dig­i­tal trans­for­ma­tion achieved sev­er­al years ago.

It turns out that killer robots that are out for our jobs or ready to take over the plan­et make effec­tive click­bait. But the unfor­tu­nate result is that the term becom­ing a watered-down ver­sion that’s stand­ing in for a num­ber of technologies.

What Is Human Intelligence?

So before we exam­ine these tech­nolo­gies, let’s go back to basics. Essen­tial­ly, the word intel­li­gence is in AI because it refers to humans. AI is the sim­u­la­tion of human intel­li­gence and our brains’ abilities.

You may think of human intel­li­gence as the abil­i­ty to sim­ply ​“know” some­thing. Of course, it’s more than that. We also have the abil­i­ty to learn, cal­cu­late, deduce, rea­son, prob­lem-solve, plan, infer, and offer an appro­pri­ate emo­tion­al response to any giv­en sit­u­a­tion. All of these ​“ways to be smart” also man­i­fest them­selves in AI.

Super Com­put­ers that Are Super Human

We know that com­put­ers can process infor­ma­tion in the form of data. That’s why we call their ​“brains” the proces­sor. But now we also want com­put­ers to hold onto the result­ing infor­ma­tion that comes from that pro­cess­ing and manip­u­late it to do more with it. AI is the activ­i­ty beyond basic pro­cess­ing, where com­put­ers start to think as well as just do. 

Our goal in the devel­op­ment of AI is not to sim­ply repli­cate and auto­mate actions that humans can per­form (although it is very much that), our wider and per­haps longer-term goal with AI is to pre­pare com­put­ers to do things that humans can’t do.

That’s not just doing things faster. We already know that com­put­ers can process deci­sions faster than human beings can. It’s about being able to han­dle com­plex deci­sions that involve increas­ing­ly sen­tient and per­cep­tive lev­els of reasoning.

The Road to Greater Knowing

For advances in AI to start prepar­ing machines to per­form more sen­tient and per­cep­tive tasks, the AI brain (the data ana­lyt­ics and pro­cess­ing engine at the heart of the AI sys­tem) needs to be exposed to as wide a vari­ety of data sources and ​“events” as pos­si­ble. This is the learn­ing part, much like humans start to learn about their envi­ron­ment as children.

As techo­pe­dia explains, ​“Knowl­edge engi­neer­ing is a core part of AI research. Machines can often act and react like humans only if they have abun­dant infor­ma­tion relat­ing to the world. Arti­fi­cial intel­li­gence must have access to objects, cat­e­gories, prop­er­ties, and rela­tions between all of them to imple­ment knowl­edge engi­neer­ing. Ini­ti­at­ing com­mon-sense, rea­son­ing, and prob­lem-solv­ing pow­er in machines is a dif­fi­cult and tedious task.”

The Tur­ing Test

British math­e­mati­cian Alan Tur­ing devel­oped his Tur­ing Test in 1950 to mea­sure a machine’s abil­i­ty to exhib­it intel­li­gent behav­ior equiv­a­lent to, or indis­tin­guish­able from, that of a human.

Tur­ing pro­posed that a human eval­u­a­tor would judge nat­ur­al lan­guage con­ver­sa­tions between a human and a machine designed to gen­er­ate human-like respons­es. His test is wide­ly respect­ed, but also wide­ly con­tro­ver­sial. This has meant that it has arguably become more of a dis­cus­sion point than an indus­try standard.

The dis­agree­ments go on: Some argue that machine learn­ing (ML) is the act of com­pu­ta­tion that builds the synaps­es inside an AI brain. Equal­ly, oth­ers argue that machine learn­ing is the result and man­i­fes­ta­tion of AI. The two terms shouldn’t be used in equal mea­sure of as switch­able replace­ments for them­selves, but they often are 

Open­ing Pandora’s Box of AI Ethics Questions

One impor­tant aspect that has sur­faced as machines get smarter is AI ethics. As we build machine brains capa­ble of rea­son­ing and mak­ing deci­sions, we need those deci­sions to be pos­i­tive ones that are good for humans and that don’t hurt or offend people.

Going deep­er here, Nat­ur­al Lan­guage Under­stand­ing (NLU), speech recog­ni­tion, syn­thet­ic speech, and Human Com­put­er Vision (HCV) are all part of AI. So we need to make sure that com­put­ers say the right things to the right peo­ple with the right lev­el of pro­fes­sion­al and cul­tur­al sen­si­tiv­i­ty. This is just one exam­ple of why AI needs some sort of frame­work for ethics. Anoth­er more-crit­i­cal exam­ple is cre­at­ing self-dri­ving cars that are able to avoid hit­ting pedestrians.

On SAP’s AI pages, there is a state­ment by Luka Mucic, chief finan­cial offi­cer at SAP who has said, ​“SAP con­sid­ers the eth­i­cal use of data a core val­ue. We want to cre­ate soft­ware that enables the intel­li­gent enter­prise and actu­al­ly improves people’s lives. Such prin­ci­ples will serve as the basis to make AI a tech­nol­o­gy that aug­ments human talent.”

Will AI Replace Humans?

We can see that AI is on the road to get­ting a whole lot smarter. But we can also see that there are a lot of dif­fer­ent ele­ments of AI. The many ways that it’s applied to our busi­ness sys­tems have yet to be refined and finessed. In its next stage of devel­op­ment, AI will become a more embed­ded and implic­it part of the tech­nolo­gies around us.

Despite some scare­mon­ger­ing here and there, it is wide­ly argued that AI will not replace human beings and the jobs that we do. Instead, it will free our time up to do more valu­able tasks that machines are still not capa­ble of doing. For now, right?

Want to bring arti­fi­cial intel­li­gence and machine learn­ing into your busi­ness? Join us on Decem­ber 11 at the NVIDIA head­quar­ters for an ASUG Exec­u­tive Exchange event.