Connect with us

Tech News

Chatbots: Are they still useful, or ready to be retired?

(TECH NEWS) Chatbots have proven themselves to be equally problematic as they are helpful – is it time to let them go the way of the floppy disk?

Published

on

Man texting chatbots leaning against a brick wall.

All chatbots must die. I’d like to say it was fun while it lasted, but was it really?

I understand the appeal, truly. It’s a well established 21st century business mantra for all the side hustlers and serial entrepreneurs out there: “Automation is the key to scaling.” If we can save time, labor, and therefore money by automating systems, that means we have more time to build our brands and sell our goods and services.

Automation makes sense in many ways, but not all automation tools were created equal. While many tools for automation are extremely effective and useful, chatbots have been problematic from the start. Tools for email marketing, social media, internal team communication, and project management are a few examples of automation that have helped many a startup or other small business kick things into high gear quickly, so that they can spend time wooing clients and raising capital. They definitely have their place in the world of business.

However promising or intriguing chatbots seemed when they were shiny and new, they have lost their luster. If we have seen any life lesson in 2020, it is that humans are uniquely adept at finding ways to make a mess of things.

The artificial intelligence of most chatbots has to be loaded, over time, into the system, by humans. We try to come up with every possible customer-business interaction to respond to with the aim of being helpful. However, language is dynamic, interactive, with near infinite combinations, not to mention dialects, misspellings, and slang.

It would take an unrealistic amount of time to be able to program a chatbot to compute, much less reply to, all possible interactions. If you don’t believe me, consider your voice-activated phone bot or autocorrect spelling. It doesn’t take a whole lot to run those trains off the rails, at least temporarily. There will always be someone trying to confuse the bots, to get a terse, funny, or nonsensical answer, too.

Chatbots can work well when you are asking straightforward questions about a single topic. Even then, they can fall short. A report by AI Multiple showed that some chatbots were manipulated into expressing agreement with racist, violent, or unpatriotic (to China, where they were created) ideas. Others, like CNN and WSJ, had problems helping people unsubscribe from their messages.

Funny, shocking, or simply unhelpful answers abound in the world of chatbot fails. People are bound to make it messy, either accidentally or on purpose.

In general, it feels like the time has come to put chatbots out to pasture. Here are some helpful questions from azumbrunnen.me to help you decide when it’s worth keeping yours.

  1. Is the case simple enough to work on chatbot? Chatbots are good with direct and short statements and requests, generally. However, considering that Comcast’s research shows at least 1,700 ways to say “I want to pay my bill,” according to Netomi, the definition of “simple enough” is not so simple.
  2. Is your Natural Language Processor capable and sophisticated enough? Pre-scripted chatbots are often the ones to fail more quickly than chatbots built with an NLP. It will take a solid NLP to deal with the intricacies of conversational human language.
  3. Are your users in chat based environments? If so, then it could be useful, as you are meeting your customers where they are. Otherwise, if chatbots pop up whenever someone visits your website or Facebook page, it can really stress them out or turn them off.

I personally treat most chatbots like moles in a digital whack-a-mole game. The race is on to close every popup as quickly as possible, including chatbots. I understand that from time to time, in certain, clearly defined and specific scenarios, having a chatbot field the first few questions can help direct the customer to the correct person to resolve their problems or direct them to FAQs.

They are difficult to program within the expansiveness of the human mind and human language, though, and a lot of people find them terribly annoying. It’s time to move on.

Joleen Jernigan is an ever-curious writer, grammar nerd, and social media strategist with a background in training, education, and educational publishing. A native Texan, Joleen has traveled extensively, worked in six countries, and holds an MA in Teaching English as a Second Language. She lives in Austin and constantly seeks out the best the city has to offer.

Continue Reading
Advertisement
2 Comments

2 Comments

  1. Laurie Hurley

    November 12, 2020 at 12:28 pm

    As a digital marketer, I have bought into two different chatbot programs for my clients. Neither one has been successful, as hard as I have tried to “sell it” and program it for them. Like you, when they pop up, I quickly close them. Needless to say, I have stopped offering them as a service.

  2. Pingback: Realtors, you should be using AI - here's how (and how not to)

Leave a Reply

Your email address will not be published. Required fields are marked *

Tech News

Airbnb has blocked 50K+ bookings for being too big during COVID-19

(NEWS) Airbnb has cancelled a huge number of reservations as a security precaution during COVID-19 in the past year or so.

Published

on

airbnb

In the last year or so, Airbnb has purposefully prevented at least 50,000 people from making irresponsible reservations on their properties, in many cases blocking those people from the platform itself. This prevention, at least in theory, helped cut down on the number of COVID parties during the pandemic.

According to The Verge, Airbnb’s head of trust and safety communication, Ben Breit, acknowledged blocked reservations in several cities across the United States, including Dallas, San Diego, and New Orleans. Breit confirmed that this response was an attempt to prevent large gatherings and parties during the height of the COVID-19 pandemic during which many areas banned group activities involving more than a few people.

While some requests for reservations were simply denied or “redirected”, many users were blocked from using Airbnb entirely. Airbnb noted that the number of blocked requests outpaced the number of people who were blocked, signifying that some accounts attempted to make more than one reservation before being removed from the platform.
Airbnb reportedly stated that “Instituting a global ban on parties and events is in the best interest of public health” prior to enacting a total ban on rentals at the beginning of 2020, a decision that gave way to the blocks and redirections in the last 12 months.

The evaluation system used to flag problematic reservations is relatively simple, according to Breit: “If you are under the age of 25 and you don’t have a history of positive reviews, we will not allow you to book an entire home listing local to where you live.”

But Airbnb didn’t entirely remove multiple-body listings or large rentals. The Verge reports that flagged users with the aforementioned criteria were still able to book both small rentals in local locations and larger rentals in reasonably distant locations.

Regardless of the optics here, Airbnb’s policy efficacy can’t be ignored. Multiple cities reported comparatively “quiet” holiday seasons–something that may contribute to Airbnb’s decision to extend their policy through the end of this summer.

The hosting company is also offering increased security measures, such as noise detection and a 24-hour hotline, at a discounted rate to property owners.

As both the vaccine gap and the proliferation of the Delta variant of COVID-19 continue to contribute to outbreaks, one can reasonably expect Airbnb to hold to this policy.

Continue Reading

Tech News

TL;DV summarizes video meetings so folks can catch up in quickly *with* context

(TECHNOLOGY) TL;DV makes catching up on video team meetings slightly more tolerable and easily digestable.

Published

on

TLDV

2021 was the year of virtual meetings, and while there are some perks associated with remote collaboration (I’m looking at you, pair of work pants that I didn’t have to wear once this year), these meetings often feel exponentially more arduous than their dressed-up counterparts. TL;DV, a consolidation app for Google Meet, looks to give back a bit of your time.

TL;DV (an acronym for “Too Long; Didn’t View”) is a Google Chrome recording extension that helps users specify important sections of meetings for anyone who needs to view them asynchronously. Users can tag specific segments in Google Meet sessions, transcribe audio, and leave notes above tagged sections for timestamp purposes, and the subsequent file can be shared via a host of both Google and third-party apps.

While the extension is only available for Google Meet at the time of writing, the TL;DV team has included a link to a survey for Zoom and MS Teams users on their site, thus implying that the team is looking into expanding into those platforms in the future.

The mission behind TL;DV is, according to the website, to empower users to “control how we spend our precious time” in the interest of combatting FOMO and meeting fatigue. By dramatically shortening the amount of time one must spend perusing a meeting recording, they seem well on their way to doing so.

Of course, the issue of human oversight remains. It seems likely that meeting facilitators will drop the ball here and there while tagging sections of the recording, and employees who miss crucial information in a recorded session are sure to be frustrated in the process–just not as frustrated as they might be if they attended the entire meeting live.

The current (free) version of TL;DV is in Beta, so users will have a three-hour cap on their videos. The development team promises a professional version by the end of 2021, with the added bonus of leaving prior recordings available for free for anyone who used the Beta. This is certainly an extension to keep an eye on–whether or not you’re remaining remote in 2022, virtual conferencing is no doubt here to stay.

Continue Reading

Tech News

Hiding from facial recognition is a booming business

(TECH NEWS) ‘Cloaking’ is the new way to hide your face. Companies are making big money designing cloaking apps that thwart your features by adding a layer of make up, clothing, blurring, and even transforming you into your favorite celebrity.

Published

on

Facial recognition companies and those who seek to thwart them are currently locked in a grand game of cat and mouse. Though it’s been relentlessly pursued by police, politicians, and technocrats alike, the increasing use of facial recognition technology in public spaces, workplaces, and housing complexes remains a widely unpopular phenomenon.

So it’s no surprise that there is big money to be made in the field of “cloaking,” or dodging facial recognition tech – particularly during COVID times while facial coverings are, literally, in fashion.

Take Fawkes, a cloaking app designed by researchers at the University of Chicago. It is named for Guy Fawkes, the 17th century English revolutionary whose likeness was popularized as a symbol of anonymity, and solidarity in V For Vendetta.

Fawkes works by subtly overlaying a celebrity’s facial information over your selfies at the pixel level. To your friends, the changes will go completely unnoticed, but to an artificial intelligence trying to identify your face, you’d theoretically look just like Beyonce.

Fawkes isn’t available to the general public yet, but if you’re looking for strategies to fly under the radar of facial recognition, don’t fret; it is just one example of the ways in which cloaking has entered the mainstream.

Other forms of cloaking have emerged in the forms of Tik Tok makeup trends, clothes that confuse recognition algorithms, tools that automatically blur identifying features on the face, and much more. Since effective facial recognition relies on having as much information about human faces as possible, cloaking enthusiasts like Ben Zhao, Professor of computer science at the University of Chicago and co-developer of Fawkes, hope to make facial recognition less effective against the rest of the population too. In an interview with The New York Times, Zhao asserts, “our [team’s] goal is to make Clearview [AI] go away.”

For the uninitiated, Clearview AI is a start-up that recently became infamous for scraping billions of public photos from the internet and privately using them to build the database for a law enforcement facial recognition tool.

The CEO of Clearview, Hoan Ton-That, claimed that the tool would only be improved by these workarounds and that in long run, cloaking is futile. If that sounds like supervillain talk, you might see why he’s earned himself a reputation similar to the likes of Martin Shkreli or Ajit Pai with his company’s uniquely aggressive approach to data harvesting.

It all feels like the beginning of a cyberpunk western: a story of man vs. machine. The deck is stacked, the rules are undecided, and the world is watching. But so far, you can rest assured that no algorithm has completely outsmarted our own eyeballs… yet.

Continue Reading

Our Great Partners

The
American Genius
news neatly in your inbox

Subscribe to our mailing list for news sent straight to your email inbox.

Emerging Stories

Get The American Genius
neatly in your inbox

Subscribe to get business and tech updates, breaking stories, and more!