INNOVATION

4 Things You Absolutely Need to Know About Software Bots

Artificial Intelligence (AI) is no longer a dis­tant fan­tasy. It’s here. Now!

Just last week, Google DeepMind’s deep learn­ing AI, AlphaGo, crushed the world’s top-ranked Go player, Ke Jie, 3-0 in a much-an­tic­i­pated ex­hi­bi­tion match at the Future of Go Summit (May 23-27).

And AI can do a whole lot more than play board games and go on TV game shows, there’s now an en­tire ecosys­tem of bots, in­clud­ing chat­bots, crawlers and trans­ac­tional bots. And as more time and money is spent on soft­ware bot de­vel­op­ment, these bots are go­ing to be­come smarter and faster, not to men­tion more di­versely skilled.

So, what can these AI pro­grams, com­monly known as bots, ac­tu­ally do? And what are the things we should know about them?

1. What Are Software Bots?

All pro­grams op­er­ate in three stages (input-processing-output), bot soft­ware are pro­grams that have a de­gree of au­ton­omy dur­ing the pro­cess­ing stage and a less lim­ited range dur­ing the in­put and out­put stages, thanks to that boost in pro­cess­ing ca­pa­bil­i­ties. Basically, AI is au­toma­tion and this au­toma­tion al­lows pro­grams to com­mu­ni­cate across a wider range of medi­ums, such as speak­ing in nat­ural lan­guage (e.g. Siri).

Bots are the soft­ware equiv­a­lents of ro­bots: au­to­mated ma­chines and au­to­mated bot soft­ware. What tasks they can at­tempt, let alone ac­com­plish, is wide-rang­ing, and varies AI to AI.

For ex­am­ple, AlphaGo and IBMs Watson are su­per­hu­man at the board game ‘Go’ and the game show ‘Jeop­ardy!′ re­spec­tively, but nei­ther can chat with you like a chat­bot can. And while chat­bots can chat, they can’t cri­tique the lat­est Hollywood block­buster, or do what AlphaGo and Watson can.

AlphaGo, Watson and chat­bots (including ad­vanced ones like Siri) are Weak AI.

Weak (Narrow) AI

A bot like AlphaGo is what is called Weak AI. But why? Isn’t AlphaGo the best player at what’s con­sid­ered the most com­pli­cated 2-player board game? Yes and yes, But while AlphaGo can play Go, it can’t play other board games, such as Chess. Its abil­i­ties are narrow’ in con­trast to its hu­man op­po­nents, who can also see, hear, taste, smell, touch, speak and much more.

Ergo: Weak AI.

A Weak AI an AI that’s non-sen­tient and only ca­pa­ble of a nar­row range of tasks, how­ever, it can be very com­pe­tent at these tasks.

The the­o­ret­i­cal al­ter­na­tive to Weak AI is Strong AI.

Strong (Deep) AI

A strong AI is an AI ca­pa­ble of a wide range of tasks.

It’s be­lieved that a Strong AI would be con­scious (or the ma­chine equiv­a­lent).

Humans have not yet been able to cre­ate a Strong AI. However, some Weak AIs are much stronger’ than oth­ers (e.g. Watson could un­der­stand ver­bal ques­tions (involving word­play) on top­ics as var­ied as lit­er­a­ture and sports, and pro­vide ver­bal an­swers phrased as ques­tions (this is how an­swers are pro­vided in Jeopardy!).

Another ap­proach to cat­e­goris­ing soft­ware bots is by intelligence’.

ANI, AGI and ASI

Artificial Narrow Intelligence (ANI) is es­sen­tially Weak AI; when the ma­chine’s in­tel­li­gence is iso­lated to a nar­row range of pa­ra­me­ters and con­texts. All ex­ist­ing AI are ANI.

Artificial General Intelligence (AGI) is much the same as Strong AI, but with stronger par­al­lels to hu­man in­tel­li­gence, which is gen­eral in­tel­li­gence. An AGI would be able to do every­thing a hu­man can do: play board games and go on TV games shows, chat, dis­cuss movies, read books and learn, be moral, etc.

Artificial Super Intelligence (ASI) is some­thing we can only spec­u­late about. Such an AI would sur­pass all hu­mans at all things: maths, writ­ing books about Orcs & Hobbits, pre­scrib­ing med­i­cine and much, much more.

Even op­ti­mistic ex­perts be­lieve AGI, let alone ASI, is decades down the line. Perhaps even cen­turies. We re­ally don’t know, and can only spec­u­late as to what an AGI or ASI would be like.

As we can see, soft­ware bots ex­ist in a com­pli­cated ecosys­tem. The weak­est of Weak AI are lit­tle more than or­di­nary pro­grams and the strongest of Strong AI would be an ASI, and ex­tra­or­di­nary.

2. What Can Bots Do?

Presently, AlphaGo plays Go, but soon Deepmind’s AI may be re­shap­ing the health­care in­dus­try along­side Watson (can­cer treat­ment). And then there’s Codebots.

Codebots is a pro­gram that cre­ates other pro­grams; code that writes code.

In soft­ware de­vel­op­ment, teams of de­vel­op­ers, de­sign­ers and pro­ject man­agers be­gin build­ing a web or mo­bile ap­pli­ca­tion by map­ping out what the ap­pli­ca­tion will be. This in­cludes writ­ing a re­quire­ments back­log.

Previously, they would then use these plans as a guide when mak­ing the pro­gram. Now they can use Codebots to trans­late di­a­grams and mod­els into code.

The core prob­lem is that al­though ap­pli­ca­tions vary in the de­tails, there are many shared arte­facts that de­vel­op­ers are re­quired to re­peat­edly re-cre­ate or adapt and im­ple­ment on an app to app ba­sis. Codebots can read and un­der­stand the high-level plans all app de­vel­op­ment nec­es­sar­ily be­gins with, and is able to write most of the code needed to bring these plans into re­al­ity in mere min­utes. On av­er­age, Codebots writes 92.68% of the re­quired back­end code.

With soft­ware bots help­ing with te­dious tasks, hu­man cre­ativ­ity has more space to thrive. Using Codebots, de­vel­op­ers, de­sign­ers, and other users are em­pow­ered to cre­ate awe­some prod­uct and ser­vice so­lu­tions (and in less time!).

In cases like Codebots, the bots are like side­kicks, with cre­ative hu­mans still the he­roes. This is also how many busi­nesses think of chat­bots, which are amaz­ing tools to have in your busi­ness’ sales and cus­tomer ser­vice toolk­its, while also be­ing a mine­field of mis­matched ex­pec­ta­tions.

Bots can write mu­sic, draw land­scapes, play Go, write code and more! And this list will ex­pand. The soft­ware de­vel­op­ment process is one where time, money and in­cre­men­tal im­prove­ment cre­ates mas­sive change. What can soft­ware bots do? Time will tell. Now is an ideas boom for AI re­search.

Read our Way of Working to find out how Codebots has evolved the soft­ware de­vel­op­ment process.

3. Are Bots Going to Take Over the World?

In his book Superintelligence, Nick Bostrom be­gins with The Unfinished Fable of the Sparrows.

Once upon a time a group of spar­rows de­cided that it would be nice if they had a big, strong owl to help them do spar­row things. When one scep­ti­cal spar­row voiced his con­cerns over how they, lowly spar­rows, could pos­si­bly con­trol an owl, an­other spar­row replied:

Taming an owl sounds like an ex­ceed­ing dif­fi­cult thing to do. It will be dif­fi­cult enough to find an owl egg. So let us start there. After we have suc­ceeded in rais­ing an owl, then we can think about tak­ing on this other chal­lenge.”

And with that, most of the spar­rows went off in search of an owl. The re­main­ing spar­rows soon re­alised that learn­ing how to tame an owl was­n’t easy, in no small part be­cause they had no owls to prac­tice on. But they pressed on as best they could be­cause, at any mo­ment, the flock might re­turn with an owl egg be­fore a so­lu­tion to the con­trol prob­lem had been found.”

Humans are the spar­rows, in Bostrom’s metaphor, and ASI is the owl. As it was for the spar­rows, the control prob­lem” is es­pe­cially chal­leng­ing be­cause we may only get one chance at solv­ing it. This is be­cause once an un­friendly su­per­in­tel­li­gence emerges, it will pre­sum­ably stymie any at­tempt we make at stop­ping it or chang­ing its pref­er­ences (e.g. the Terminator se­ries’ Skynet). (That’s why our own Codebots tech­nol­ogy can only cre­ate and edit eter­nal code, and not its own, in­ter­nal code.)

But should we be wor­ried about AI tak­ing over the world?

Well, it de­pends. Bots may be beat­ing us at board games and TV game shows, but soft­ware bot busi­nesses still have a long way to go in de­vel­op­ing AGI, let alone ASI.

Even emerg­ing tech­nolo­gies like Codebots re­quire hu­man pi­lots.

Still wor­ried?

Your brain is an im­mense net­work of about 100 bil­lion neu­rons con­nected by around 100 tril­lion synapses. You have hun­dreds of tril­lions of neural path­ways in your head. No com­puter in the world to­day comes close to match­ing your brain’s horse­power.

The Fujitsu-built K, one of the world’s fastest su­per­com­put­ers, is ca­pa­ble of 8 quadrillion cal­cu­la­tions per sec­ond, and it took 40 min­utes to sim­u­late a sec­ond of neural ac­tiv­ity!

Skynet is­n’t even re­motely pos­si­ble with to­day’s tech­nol­ogy. So no, even the most lead­ing edge bots are not go­ing to take over the world, but they may take your job.

4. Are Bots Going to Take My Job?

Many in­dus­tries are cur­rently be­ing dis­rupted by emerg­ing AI tech­nolo­gies, and many more may soon be.

According to PricewaterhouseCoopers, 38% of US jobs could be au­to­mated by the early 2030s, with sim­i­lar per­cent­ages ex­pected in Germany (35%) and the UK (30%).

The threat of au­toma­tion is high­est in the trans­porta­tion and stor­age (56%), man­u­fac­tur­ing (46%) and whole­sale & re­tail (44%) sec­tors, and lower in health and so­cial work (17%).

Workers with un­der­grad­u­ate and post­grad­u­ate de­grees have a much lower risk of be­ing au­to­mated.

But it’s not all doom and gloom. Fresh de­vel­op­ment into AI and new au­toma­tion tech­nolo­gies are ex­pected to cre­ate new in­dus­tries and re­vi­talise oth­ers.

Summary

When it comes to the AI pro­grams com­monly known as bots, there’s a lot to know. For one, bots ex­ist in a com­pli­cated ecosys­tem, with the weak­est of Weak AI be­ing lit­tle more than or­di­nary pro­grams, while the strongest of Strong AI are ASI, and truly ex­tra­or­di­nary.

This ecosys­tem is fur­ther mud­dled by bots such as Siri, which are more a col­lec­tion of sep­a­rate AI processes than a sin­gle co­her­ent unit.

And then there’s the threat posed by these ma­chines. Automation could see hu­man work­ers re­placed by ro­bots and bots across a di­verse spec­trum of sec­tors and in­dus­tries (luckily some­thing like Skynet is­n’t even re­motely on the hori­zon).

What’s clear is that the ro­bots are ris­ing, and we had best un­der­stand what soft­ware bots are if we are to make our way in this new, in­ter­est­ing world.

Contact us here to dis­cuss your soft­ware pro­ject fur­ther.

ABOUT THE AUTHOR

Matt Francis

Brewer of beers, smoker of meats

Get cu­rated con­tent on soft­ware de­vel­op­ment, straight to your in­box.

Your vi­sion,

our ex­per­tise

Book an con­sul­ta­tion