Robert Desprez Communications

ChatGPT: The AI-Powered Proofreader

by Leave a Comment

Chat­G­PT, a text-gen­er­at­ing mod­el devel­oped by Ope­nAI to sim­u­late respons­es that resem­ble human con­ver­sa­tions, saves you time and effort by edit­ing your documents.

Copy a page of con­tent into Chat­G­PT and request that it sum­ma­rize the infor­ma­tion. In sec­onds, it sum­ma­rizes the mate­r­i­al before your eyes—much faster than any human that I know.

For a tool that is just emerg­ing, it is sur­pris­ing­ly ver­sa­tile. You can ask it for help with pro­gram­ming, brain­storm­ing, writ­ing head­lines (Chat­G­PT sug­gest­ed the head­line for this blog post), trans­la­tions, mar­ket­ing mate­r­i­al, and plan­ning a trip.

A quick simple test

I decid­ed to give it a test. I copied an ear­li­er blog that I draft­ed about Mad­Cap Flare and delib­er­ate­ly added errors (high­light­ed in red) to the text. In my state­ment to Chat­G­PT, oth­er­wise known as a prompt, I wrote, “Check the fol­low­ing con­tent for incon­sis­tent acronyms, con­sis­tent spelling of com­pa­ny names, gram­mar and typos, and make sure all num­bers under 10 are spelled out.” 

In under 10 sec­onds, the AI tool gen­er­at­ed a new ver­sion of the con­tent and sum­ma­rized the edits made:

Chat­G­PT quick­ly caught all the errors and even cor­rect­ed the case of “Robo­Help,” which is the offi­cial name of the help author­ing tool.

Purge passive voice using ChatGPT

Many tech­ni­cal writ­ing depart­ments fer­ret out exam­ples of pas­sive voice and strive to re-write the con­tent so it is active because the writ­ing is more suc­cinct and eas­i­er to under­stand (of course, there is a place for pas­sive voice). In the fol­low­ing exam­ple, I asked Chat­G­PT to revise the fol­low­ing sen­tences that use pas­sive voice:

In moments, Chat­G­PT revised the sen­tences so they were active:

ChatGPT does have limitations

The AI tool does have its limitations:

  • Lacks under­stand­ing: If you write a tech­ni­cal report about a mine or a pro­pri­etary soft­ware pack­age, it will not under­stand the details of the con­tent. Sub­ject Mat­ter Experts must still vet the accu­ra­cy of the content.
  • Good prompts are required: Typ­ing vague com­ments like “edit this doc­u­ment” and the tool may not edit what you require. If you want to check a doc­u­ment for pas­sive voice, include the specifics of what you need.
  • Chat­G­PT was last updat­ed in 2021: If you require cur­rent infor­ma­tion about a major world event last year (such as Rus­sian’s inva­sion of Ukraine), Chat­G­PT will be “in the dark.”
  • Be care­ful of con­fi­den­tial­i­ty. See Gen­er­a­tive AI Chat­G­PT Can Dis­turbing­ly Gob­ble Up Your Pri­vate And Con­fi­den­tial Data, Fore­warns AI Ethics And AI Law

Still, it is sur­pris­ing­ly pow­er­ful and fast tool that can help with writ­ing and edit­ing. Check out ChatGPT. 

Do We Rely Too Much on Technology?

by Leave a Comment

Does soft­ware automa­tion improve or ham­per your life?

Most peo­ple embrace the idea of tech­nol­o­gy. How many peo­ple real­ly would rel­ish writ­ing a let­ter or even an email with­out spell check? Or endure a rush-hour dri­ve into a for­eign city with­out a GPS system?

Nicholas Carr, author of The Glass Cage, argues that for all of technology’s ben­e­fits, it has its share of per­ils. When peo­ple use com­put­ers, they often fall vic­tim to a pair of cog­ni­tive ailments—automation com­pla­cen­cy and automa­tion bias.

Technology Can Harm our Careers and Threaten Lives

Automa­tion com­pla­cen­cy lulls us into a false sense of secu­ri­ty. Carr writes: “We become so con­fi­dent that the machine will work flaw­less­ly, han­dling any chal­lenge that may arise, that we allow our atten­tion to drift. We dis­en­gage from our work, or at least the part of it that the soft­ware is han­dling, and as a result may miss sig­nals that some­thing is amiss.” Any­one who uses a word proces­sor will have expe­ri­enced this: per­haps the spell check doesn’t catch an embar­rass­ing error. Or a foot­er does not dis­play the cor­rect information.

Automa­tion bias occurs when peo­ple give too much cre­dence to the infor­ma­tion dis­played by a com­put­er. Even when the infor­ma­tion is wrong or mis­lead­ing, peo­ple may still believe it. One alarm­ing exam­ple is a dri­ver of a school bus who mis­tak­en­ly rammed into a con­crete bridge in Seat­tle, injur­ing 21 stu­dents. The dri­ver told police that he had been fol­low­ing the GPS instruc­tions and “did not see” signs and flash­ing lights warn­ing of the impend­ing bridge.

Soft­ware automa­tion is every­where in our soci­ety. Accoun­tants use deci­sion-sup­port soft­ware in cor­po­rate audits. The soft­ware expe­dites the work but “there are trou­bling signs that as the soft­ware becomes more capa­ble the accoun­tants become less so,” writes Carr. One study, con­duct­ed by a group of Aus­tralian pro­fes­sors, exam­ined the effect of the soft­ware sys­tems used by three account­ing firms. Those accoun­tants who had to make more of the deci­sions them­selves had “a sig­nif­i­cant­ly stronger under­stand­ing of dif­fer­ent forms of risk” than the oth­er accoun­tants who relied on the soft­ware. “The decline in learn­ing asso­ci­at­ed with advanced soft­ware affect­ed even vet­er­an audi­tors,” writes Carr.

A ten­den­cy to be increas­ing­ly reliant on soft­ware is not unique to accoun­tants. Finan­cial pro­fes­sion­als, lawyers, doc­tors, human resource pro­fes­sion­als, pilots all rely on soft­ware to com­plete aspects of their jobs. The book con­tains many exam­ples of how these well-edu­cat­ed pro­fes­sion­als become over­ly reliant on tech­nol­o­gy and make mistakes—sometimes life-threat­en­ing ones.

Online Information May Weaken our Memory

In the devel­oped world, most of can­not imag­ine our lives with­out an Inter­net con­nec­tion and a search engine.

When you type in a search query using Google, the search engine does a great job of fig­ur­ing out what you want. It cor­rects your spelling. It sug­gests search terms. It even antic­i­pates your needs based on who you are and the terms you have used in the past. Over time, you might think that users are learn­ing from Google’s sug­ges­tions so that they are more savvy when typ­ing search terms.

The oppo­site is true. A series of exper­i­ments in Sci­ence mag­a­zine show that all the avail­able online infor­ma­tion weak­ens our mem­o­ry for facts. Why remem­ber fac­tu­al con­tent when a search engine will do it for you? The researchers con­clud­ed: “Since search engines are con­tin­u­al­ly avail­able to us, we may often be in a state of not feel­ing we need to encode the infor­ma­tion internally.”

Amit Sing­hal, Google’s lead engi­neer, notes that the bet­ter Google’s search engine becomes, the lazier we are with our typ­ing and queries.

Most of us wouldn’t accept a world with­out soft­ware, whether it’s a GPS, bank­ing soft­ware, or a smart­phone. Tech­nol­o­gy has rev­o­lu­tion­ized our lives in many pos­i­tive ways. But for those of us enam­oured with tech­nol­o­gy, Carr’s book is a a cau­tion­ary tale—don’t let your guard down too much when using it.

Improve your training with practice and feedback

by Leave a Comment

Have you ever attend­ed a course or train­ing ses­sion that missed its mark?

Most of the cours­es that I have tak­en haven’t met my expec­ta­tions for dif­fer­ent rea­sons. In uni­ver­si­ty, there was the pro­fes­sor who’d lec­ture for an hour, lulling stu­dents to sleep. There was the work­place train­er who focused on how the soft­ware worked but not the tasks that stu­dents (or learn­ers) required. There was the employ­ee who was threat­ened about shar­ing too much for fear of los­ing her job so she with­held knowl­edge, ren­der­ing the train­ing almost use­less. I could go on.

In my expe­ri­ence, senior man­age­ment holds the belief that any­one can train. But few peo­ple do it well.

Too often train­ing is just a “pletho­ra of regur­gi­tat­ed knowl­edge,” said Jim Tall­man, pres­i­dent of North Pacif­ic Train­ing & Per­for­mance Inc. Tall­man spoke at the Cana­da West Chap­ter of the Soci­ety for Tech­ni­cal Com­mu­ni­ca­tion for a pro­fes­sion­al devel­op­ment ses­sion on Sep­tem­ber 13th.

Train­ing that focus­es on rehash­ing knowl­edge doesn’t help learn­ers per­form their jobs bet­ter and pro­vides incon­sis­tent results, he said.  Some exam­ples of inef­fec­tive train­ing include:

  • The cur­ricu­lum fol­lows the­o­ry, not skills or tasks that one per­forms on the job.
  • Con­tent fol­lows course design, not a task analysis.
  • Exam­ples focus on sys­tem fea­tures, not work­place processes.
  • The course high­lights the knowl­edge of the sys­tem, not improved performance.
  • The instruc­tor nev­er stops talking.

A better way to train

An alter­na­tive method of train­ing focus­es on improv­ing the learn­ers’ skills. Some ele­ments of effec­tive training:

  • Task analy­sis dri­ves course design.
  • Cur­ricu­lum reflects the job.
  • Con­tent is prac­tice-based and learn­er paced.
  • Course results in improved job performance.

He said: “This is the pre­ferred way…that train­ing should be designed to give learn­ers what they need and clients their money’s worth. Instruc­tion is based on how learn­ers learn and not on how much the train­er can demo his or her prowess in the field.”

Train­ing should be ide­al­ly 30 per­cent pre­sen­ta­tion or lec­ture and 60 per­cent hands-on (such as exer­cis­es or activ­i­ties), said Tall­man. The final 10 per­cent of the course should be allo­cat­ed on feed­back so that learn­ers can improve. Most tra­di­tion­al train­ing is reversed with the instruc­tor talk­ing most of the time and learn­ers’ eyes glaz­ing over after 20 min­utes of lec­ture, he said.

In addi­tion, effec­tive train­ing should be also mean­ing­ful, mem­o­rable, moti­va­tion­al, and mea­sur­able, Tall­man said.

  • Mean­ing­ful: Train­ing should be rel­e­vant for learn­ers so they can per­form key tasks rel­e­vant to their jobs.
  • Mem­o­rable: Learn­ers need to remem­ber their train­ing after the course. Job aids, hand­outs, and quick ref­er­ence cards may be able to assist.
  • Moti­va­tion­al: Train­ing should moti­vate learn­ers in the class­room so that they want to apply the new skills.
  • Mea­sur­able: Effec­tive train­ing teach­es new skills that can be observed. “Good learn­ing events stress observ­able behav­iors, give learn­ers suf­fi­cient prac­tice over time, and pro­vide per­for­mance feedback.”

Not sur­pris­ing­ly, Tall­man’s train­ing focused on prac­tice and feed­back so that learn­ers could improve. This was an excel­lent presentation!

Managing risk in your career

by Leave a Comment

Many Cana­di­ans are not risk tak­ers. We are a nation of home buy­ers (homes are per­ceived as safe and secure by many) and we grav­i­tate towards sta­ble jobs, ide­al­ly with full ben­e­fits and pensions.

In the book Start-up of You, LinkedIn co-founder and author Reid Hoff­man argues that overt­ly pur­su­ing safe­ty and secu­ri­ty in your career can leave you “…exposed to a huge blow-up in the future.”

Think of gov­ern­ment work­ers. They typ­i­cal­ly don’t take huge risks along the way in their careers. For their loy­al­ty, they receive incre­men­tal pay increas­es, good ben­e­fits, and a gen­er­ous pen­sion. For many peo­ple, this is utopia.

Com­pare these gov­ern­ment work­ers to real-estate agents, who need to hus­tle to find their work. Occa­sion­al­ly, they sell an expen­sive home and col­lect a lucra­tive com­mis­sion. The next month, how­ev­er, they may not sell any­thing. Their incomes are great some times, fol­lowed by lean periods.

Hoff­man writes: “He [the gov­ern­ment work­er] always eats well…until the day comes that gov­ern­ment pen­sions explode or aus­ter­i­ty mea­sures wipe out his depart­ment. Now he’s screwed. He will starve because, unlike the real estate agent, he has no idea how to deal with the downs…” A cen­tral theme of Hoff­man’s chap­ter on risk is that work­ers should take cal­cu­lat­ed risks in their careers to make them­selves more resilient when some­thing unex­pect­ed happens.

Title wave

Peo­ple don’t just per­ceive gov­ern­ment to be low risk. In the late 1990s, Nor­tel could do no wrong—it was seen as a great place to work and its CEO was her­ald­ed as a cap­i­tal­ist hero. At the time, employ­ees were envied. A few short years lat­er, after the dot-com bub­ble burst, Nor­tel laid off 60,000 employ­ees. In 2009, the com­pa­ny filed for bank­rupt­cy and now its pen­sion is in tat­ters. Long-term employ­ees must have been stunned.

Taking intelligent risks

If you lived on an African savan­nah in cen­turies past, avoid­ing risk—like com­ing face to face with a lion—made per­fect sense. Today, in the devel­oped world, most of us don’t need to wor­ry about get­ting eat­en alive. Hoff­man writes: “Over­es­ti­mat­ing risks and avoid­ing loss­es is a fine strat­e­gy for sur­viv­ing dan­ger­ous envi­ron­ments, but not for thriv­ing in a mod­ern career. When risks aren’t life-threat­en­ing, you have to over­come your brain’s dis­po­si­tion to avoid sur­viv­able risks.”

Hoff­man does not advo­cate tak­ing stu­pid risks. He argues that tak­ing risks makes sense when there’s a promise of a reward.

Here are some exam­ples of tak­ing risks in your career:

Risk Pos­si­ble reward
Work­ing on contract. Earn more income.
Improve the diver­si­ty of your work.
Increase your network.
Mov­ing to anoth­er country. Immerse your­self in anoth­er culture.
Learn anoth­er language.
Increase your net­work abroad.
Learn­ing on the job, even if it for less money. Learn new tech­nolo­gies or skills could that open doors down the road.
Return­ing to school. Learn new skills and knowledge.
Increase your net­work with peers.

Oth­er than con­tract­ing, I some­times take a cal­cu­lat­ed risk by pre­sent­ing a top­ic at an upcom­ing tech­ni­cal writ­ing con­fer­ence. How is that a risk? Because I don’t always know every­thing about the sub­ject. In some cas­es, I’ve just scratched the sur­face of a top­ic and I need to research the con­tent, cre­ate and prac­tice the pre­sen­ta­tion, and deliv­er it to a crowd of peers in a rel­a­tive­ly short time. This forces me to learn a lot about a sub­ject quick­ly and there is always a pos­si­bil­i­ty of fail­ure. Of course, there are ben­e­fits too. I learn new skills and I present my new knowl­edge to a crowd, which increas­es my vis­i­bil­i­ty. I also can use that research to present at oth­er events and writes arti­cles for mag­a­zines, such as the Soci­ety for Tech­ni­cal Com­mu­ni­ca­tion’s Inter­com magazine.

In the short term, avoid­ing risk seems like a good path for­ward. But over the long-haul, Hoff­man writes “low volatil­i­ty leads to increased vul­ner­a­bil­i­ty because it ren­ders the sys­tem [and your career] less resilient to unthink­able exter­nal shocks…If you don’t find risk, risk will find you.”

Do you agree? If so, how do you take risks at your work?

Are we facing a “skills gap” in Information Technology?

by 12 Comments

What new skills have you learned on your own time in the last six months?

If you’re like many IT peo­ple, you may be defer­ring to your boss when it comes to career devel­op­ment. Per­haps your com­pa­ny is ready to imple­ment a new Share­Point site and you’ve been immersed in that project for the last six months. Great. But have you been keep­ing your eye on indus­try trends on what skills are emerg­ing and those that are wan­ing? I sus­pect most peo­ple work­ing it IT would say “no.”

The June 4th edi­tion of Com­put­er­World mag­a­zine argues that com­pa­nies have lots of job open­ings but there are still many IT work­ers who are strug­gling to land jobs. The rea­son? A skills gap. The jobs are there but work­ers’ skills are not.

Todd Wein­man, a recruiter quot­ed in the arti­cle, said: “You can’t rely on a com­pa­ny for your growth and train­ing anymore…Except for a few enlight­ened com­pa­nies, if they’re train­ing you at all, they’re train­ing you for what they need, not nec­es­sar­i­ly train­ing for what you need to devel­op your tech­ni­cal skills over the long run.”

Employ­ers are clam­or­ing for can­di­dates with exper­tise in mobile app devel­op­ment, cloud com­put­ing, and busi­ness ana­lyt­ics, accord­ing to the arti­cle. But employ­ees may be tired of the self-devel­op­ment treadmill—especially if they are work­ing more than 50 hours per week.

Accord­ing to the arti­cle, here are some sug­ges­tions for thriv­ing in IT in the com­ing years:

* Cre­ate your own train­ing plan. Accord­ing a recent arti­cle pub­lished by the Har­vard Busi­ness Review, set aside a few hours each week on career development.

* Pur­sue train­ing and cer­ti­fi­ca­tions on your own time and at your expense. On a relat­ed note, I recent­ly read a book called the Six Fun­da­men­tals of Suc­cess by Stu­art R. Levine. Levine writes: “Take con­trol of your con­tin­u­ing edu­ca­tion and per­son­al growth.  Every six months, give your­self spe­cif­ic learn­ing goals…Too often peo­ple wait for their boss to tell them what skills they need. This is a stun­ning abdi­ca­tion of responsibility.”

* Devel­op soft skills, like com­mu­ni­ca­tion.

* Iden­ti­fy dif­fer­ent class­es of skills. Which tech­nolo­gies are emerg­ing, main­stream, and or lega­cy? Does your cur­rent exper­tise have a long future ahead of it? If not, what cer­ti­fi­ca­tions and train­ing do you need?

Here’s my take on it. I’ve worked about 15 years in IT and work­ers who enjoy learn­ing and are moti­vat­ed to learn appear to do bet­ter in the work­place. In con­trast, those employ­ees who decide to go on “career cruise con­trol” become vul­ner­a­ble dur­ing lay­offs because their skills are per­ceived to be “rusty.”

Obvi­ous advice? Per­haps. But many work­ers don’t seem to fol­low it. What do you think about this issue?

Check out the Com­put­er­World arti­cle.

 

Recent Blog Posts

About Me

Robert Desprez I have worked as a Vancouver technical writer for more than 20 years, working at Kodak, Boeing, Teck Resources, and FortisBC. In addition, I have worked as a writing instructor at Simon Fraser University in Vancouver.

Contact Me

Robert Desprez Com­mu­ni­ca­tions Inc.
North Van­cou­ver, British Columbia
Canada
Phone: 604–836-4290