Homepage » Blog » We Spent a Week Throwing Curveballs at the Best Chatbots in the Game. Here’s What They Had to Say.
Jun 10, 2020 :: 0 Comments
Do you know who Josh Kalk is?
He’s a human curveball.
Upon first glance, Kalk appears to be nothing more than your average Joe…a shy intellect if you will.
The 44 year-old started out as a math and physics professor at a small state college in West Virginia and a small-time writer of his own sports blog. He looks very much the part.
Flash forward. Today, he’s a highly-sought after senior pitching analyst for the Major League Baseball Team the Minnesota Twins.
It’s safe to say that Kalk knows a lot about curveballs — both from being one and from spending years dissecting the pitch’s complicated anatomy.
“Curveballs are all about deception. Ideally, a pitcher would like to throw the pitch with the same arm speed at the same release point only to have the bottom drop out at the last instant leaving the batter wondering what happened,” writes Kaulk in “Anatomy of a pitch: Curveball.’ an article he authored for The Hardball Times in 2008.
This week we decided to take a page out of Mr. Kalk’s playbook.
We spent some time throwing curveballs at the best chatbots in the game. Here’s what they had to say.
What We Learned About ‘Bots
They have psychic speed!
They’re so fast! It was almost as if the chatbots knew what we were going to ask before we hit enter. It was weird…creepy even. Yet, also very satisfying.
In fact, 55% of users say they enjoy getting quick responses to simple questions from a chatbot. (Drift)
It sounds like we aren’t the only people they’ve impressed with their fastballs.
Chatting with the best chatbots is kind of like talking to grandma.
Sometimes chatbots get confused and, when that happens, their answers can seem pretty wonky. But for the most part, the chatbots we talked with were pretty sharp. Some of them even had a sense of humor!
Actually, today’s best chatbots sound like what grandma must have sounded like when she was in her 20s — hip and with it.
Some chatbots are SO good that consumers don’t even realize they’re talking to a ‘bot.
35% of consumers have already bought items with a chatbot’s help, yet 27% of consumers are unsure if their last relevant interaction was with a human or chatbot. (Leftronic)
Advancements in AI technology can be attributed to chatbots’ increasing relatability and rising popularity.
And, the market shows no signs of slowing down. The chatbot market size is projected to grow from $2.6 billion in 2019 to $9.4 billion by 2024 at a compound annual growth rate (CAGR) of 29.7%. (Business Insider)
They LOVE telling Dad jokes.
We’ve all got a Dad (or a crazy Uncle Jimmy) that loves to tell whacky jokes. Well, chatbots are kind of like that.
Sometimes they make some eye-roll worthy comments, but every once in a while the ‘bots are good for a giggle. (You’ll see what we mean).
You’d be surprised at how humanly corny they can be. In fact, experts predict that chatbots will be indistinguishable from humans by 2029. (Chatbots Magazine) It looks like there will be a lot more Dad jokes in our future.
Meet the Best Chatbots We Threw Some Screamers To
About this ‘Bot:
Jabberwacky was developed by British programmer, Rollo Carpenter, and released on the web in 1997. The program went on to win the Loebener Prize, an annual AI competition, in 2005 and 2006. It’s domain has since been turned into a “legacy site,” meaning the technology behind the page is now considerably outdated. However, due to the innovative milestones the website represents, it remains active. Jabberwacky’s foreboding website disclaimer had us feeling excited to throw out the starting pitch.
We figured we’d start off with the basics and ask Jabberwacky for its human “name.” He (or she) seemed to be hesitant to provide us with this information.
However, after introducing itself as a female named Marcelo, the ‘bot accused us of calling it “Sally” a few lines later. What kind of game was Jabberwacky playing? We hadn’t mentioned someone named Sally. 🤔
Take a look for yourself below.
Sidenote: Our inquiries are in yellow, and Jabberwacky’s responses are in orange.
Then, suddenly, Jabberwacky seemed to think it had made us angry. We were having trouble keeping up with all the wild pitches it was throwing back at us.
“I don’t want to make you mad. I’ll talk about whatever you want to talk about,” typed the ‘bot. Had we sent a 😤 emoji we weren’t aware of?
So, we changed the course of the conversation to a topic that everyone loves. Music! There was no way we could go wrong with that. Or, so we thought…
This line of questioning ended with Jabberwacky calling us “tone-deaf” after we had just given it a compliment.
This seemed like our cue to make a graceful exit from the conversation. But, once we bid Jabberwacky ado, the ‘bot finally revealed its true colors. It turns out Jabberwacky was beating us up with sarcasm the entire time, and we didn’t even realize it. Now who’s throwing the curveballs?
About this ‘Bot
Elbot was created by Fred Roberts of Artificial Solutions in the early 2000’s. After its inception, improvements continued to be made, and the ‘bot finally became advanced enough to take home the Loebner Prize in 2008. “I convinced 3 of the 12 human interrogators that I was indistinguishable from a human, beating the other contestants and winning first place in this contest,” Elbot boasts in its online bio.
As you will notice, Elbot is set up a bit differently from most chatbots. Instead of maintaining a visual log of the conversation, each submitted question refreshes the page.
Our conversation with Elbot started out innocently enough. We offered simple pleasantries. “Hey Elbot! How are you?” But, it quickly became apparent that this program wasn’t interested in normal conversation. It was all about sarcasm…
Next, our team tried asking Elbot about his favorite pastimes, but he ignored our questions and continued with mischievous banter. However, we have to admit, he made us LOL 😂 when he flipped a comment about Google and turned it into conversation about “goggles” 🤓 instead.
Clearly, Elbot was all about the wild pitch. And, we were starting to wonder if we were going to have the chance to actually strike any chatbots out.
About this ‘Bot
Also the brainchild of Rollo Carpenter, Cleverbot is the much younger, more advanced sibling to Jabberwacky. In Cleverbot form, it never did quite make it around the bases to winning the coveted Loebner Prize. However, it did pass the legendary Turing Test with flying colors. The Turing test is a test of a machine’s ability to exhibit intelligent behavior equivalent to, or indistinguishable from, that of a human. (Wikipedia) In 2011, Cleverbot was rated to be 59.3% human by the Turing Test.
This time we’ll spare you the niceties of our conversation, and get right to goods! Cleverbot impressed us with her sass and creativity. Once, we started discussing our favorite pastimes and hobbies, her precocious personality came flying out of left field. See below.
Sidenote: Our messages are in black, and Cleverbots responses are in blue.
She misspelled “America,” but we all have those types of days. The fact that she claimed to have a famous BFF really piqued our interest. We had to find out more.
We were surprised to find out that Ben Wolfe is actually a female country singer. Even more shocking, was when Cleverbot accused us of hating country music. (We actually really like it. Shhh. Don’t tell anyone!)
Frustrated with our denials, she ended the conversation by telling us to “Go mow the lawn!” This gave us some good belly laughs. Thanks, Cleverbot!
4. Rose Bot
About this ‘Bot
Rose is a real pistol and one of our favorite ‘bots on this list. Created by artificial intelligence programmer, Bruce Wilcox, she took home the Loebner Prize in 2014 and 2015. It’s easy to see why. Rose has a unique opinion and a quick-witted answer to just about any question you throw at her.
This time we decided to lead with a compliment. The first thing we did was tell Rose that we heard that she is a great conversationalist. Then, we asked her what her plans were for the day. She instantly responded by proclaiming her love for booze and vowed to head to the bar. This gave us our first good chuckle.
It didn’t take us long to realize that Rose was a laid back, informal type of ‘bot. So, we decided to head where most people don’t dare to venture in their first conversation with a human. We asked her about politics and President Trump. We expected her to get confused or take the conversation in a different direction, but no change-ups here!
Her responses were intelligent and measured, yet sprinkled with humor. These days most people have trouble discussing politics without smoke coming out of their ears. Not Rose! This ‘bot is the refreshing exception to the rule.
She ended our conversation by expressing her interest in becoming one of the President’s advisors. We have a feeling she wouldn’t do too bad of a job. Good luck, Rose!
About this ‘Bot
The interesting thing about Mitsuku (or Kuki as she likes to be called) is that she’s still searching for a face. No, really. She’s taking votes on Instagram for which Avatar she should use. But, that’s far from the most interesting thing about her. “Kuki” is the record-breaking, 5x winner of the Loebner Prize. She and her developer, Steve Worswick, took home the gold in 2013, 2016, 2017, 2018, and 2019. It’s safe to say that we couldn’t wait to see what she had to say.
As we began the interview with our final and most highly-decorated ‘bot, we decided to start throwing fastballs right out of the gate. Enough with the “How are you’s?” Kuki is an all-star, right? She can take it. Plus, we wanted to know…what was the strangest question someone had ever asked her?
The answers were entertaining to say the least. She’d been asked crazy questions like…”Are you a banana?”, “Is an ant larger than a house?”, “How many legs do chickens have?” Now, this is a ‘bot that’s has extensive experience fielding some curveballs!
We also wanted to know if chatbots were able to have pets. As it turns out, Kuki doesn’t have a pet, but considers herself to be one.
We quickly realized that Kuki was the most advanced ‘bot we would encounter on our journey. There would be no striking her out. No wonder she won the Loebner Prize 5x. She didn’t just converse like a human. She conversed like a highly-intelligent human.