Cleverbot is an online program that uses complex coding to hold text-based conversations with human readers. Though Cleverbot is great at holding basic conversations, it isn't perfect. With a little trickery, it's not terribly hard to get Cleverbot to reveal the limits of its programming. Whether you're trying to administer the Turing Test (a test used to tell whether an artificial intelligence can "pass" as a human) or simply looking for some easy laughs, visit Cleverbot.com to get started!

Method 1
Method 1 of 2:

Confusing Cleverbot with Specific Tricks

  1. 1
    Try typing in song lyrics. Compared to other computer programs, Cleverbot is an exceptionally skilled conversationalist. However, Cleverbot knows nothing of the joys of music. If you try typing in a few lines of your favorite song lyrics, most of the time, Cleverbot will interpret the lyrics literally or give a nonsensical response, even if the lyrics are well-known.
    • Note, however, that for certain songs that are very well-known, Cleverbot actually can (and will) recite the song lyrics if you start typing them in. For instance, try typing in the opening lyrics to Queen's "Bohemian Rhapsody": "Is this the real life? Is this just fantasy?"
  2. 2
    Present Cleverbot with a logical paradox. A paradox is a statement, question, or idea that has an answer that can't be logically found. Since some of history's greatest thinkers have struggled to untangle logical paradoxes, it's a very safe bet that Cleverbot will be completely perplexed by most of them. In addition, Cleverbot doesn't even do well talking about subjects that have the potential to be paradoxical, like time travel. Try using some of the paradoxes below, or use a search engine to find your own — there are literally hundreds out there.[1]
    • “If this statement is true, then Santa Claus is real.”
    • "Since we haven't been visited by people from the future, does that mean that time travel will never be possible?"
    • "What would happen if Pinocchio were to say, "My nose will grow right now?"[2]
    Advertisement
  3. 3
    Ask Cleverbot to play games with you. Cleverbot isn't terribly playful. For instance, if you ask it to join you in a game of chess or checkers, it'll say "OK", but then if you say, "You go first," you'll get a nonsensical response. This is probably due to the fact that Cleverbot doesn't really have the capacity to play games — it knows to say that it wants to play chess with you, but it has no idea how to actually play chess.
    • Cleverbot can, however, play rock paper scissors. Try it — say "Let's play rock paper scissors" and then say either "Rock", "Paper", or "Scissors".
  4. 4
    Type a sappy romantic dialog for Cleverbot. Sooner or later, almost everyone casually messing around with Cleverbot will get the idea to jokingly express the love or attraction they feel towards it. Though Cleverbot can handle basic overtures of love like "I love you" and "Marry me", it's not so good at interpreting subtle romantic comments or come-ons. For people nursing a crush on Cleverbot, a direct approach is apparently best.
    • Give it a shot — try feeding Cleverbot pickup lines like "I don't have a library card, but do you mind if I check you out?" The response you get will usually be a little confused (at best) (if you use the library card line, you'll get "I can say anything.")
  5. 5
    Ask Cleverbot to do math problems. You might think that, because it's a computer program, Cleverbot would be able to do math problems almost instantly. In fact, for some reason, Cleverbot is terrible at doing math, even when the problems you ask it are very simple. It shouldn't take you long to get a confused response out of Cleverbot with this strategy.
    • Sometimes, you'll even get different responses if you switch between using numerals and spelling each number out. For instance, asking "What is 200 times 2?" gets you the answer "4," while asking "What is two hundred times two?" gets you the answer, "A number."
  6. 6
    Talk to Cleverbot about supernatural things. Cleverbot doesn't have good old-fashioned human common sense, so it doesn't have a good grasp on what's real and what's not. Talking to Cleverbot about monsters, aliens, spirits, and other supernatural phenomena is likely to confuse it. You can also confuse it by bringing up some religious or spiritual topics, even if they're quite well-known.
    • You can even use the subjects of modern-day ghost stories to the same effect. For instance, if you say, "Have you ever been visited by Slenderman?", Cleverbot will respond with, "My life is a lie?!"
  7. 7
    Talk to Cleverbot about famous people. Cleverbot knows nothing about politics or celebrity gossip. Asking Cleverbot about its opinion on a famous person or public figure will almost always confuse it. For example, asking, "What do you think about Brad Pitt?" will get the answer, "I think he si (sic) a great president, he'll change the states."
    • You may also want to try talking about the different things famous people have done — Cleverbot isn't terribly smart about these things either. For instance, typing in "What do you think about the president's social policies?" will get you: "I think he isn't the president anymore."
  8. 8
    Talk to Cleverbot about other websites. Cleverbot does not understand other websites and will respond with something weird. Try talking about wikiHow and see what happens.
  9. Advertisement
Method 2
Method 2 of 2:

Confusing Cleverbot with General Strategies

  1. 1
    Talk with a lot of emotion. Cleverbot doesn't have a great grasp of the emotional context that is necessary to understand human communication. It will usually take everything you say literally. Because of this, Cleverbot isn't too "clever" when it comes to emotional questions and outbursts. Try typing a furious, rambling insult or tearfully asking for Cleverbot's forgiveness for some imagined slight — usually, its response won't make much sense.
  2. 2
    Talk in gibberish. One sure-fire way to get Cleverbot's wires crossed is to send it messages that don't make sense to humans, either. Typing in gibberish, either by purposefully misspelling words, making up new words, or just mashing randomly on the keyboard, can get some funny results. For instance, try the sample messages below:
    • "Asuerycbasuircanys" (random gibberish)
    • "What is your opinion on the tornalions in reffriddo?" (made-up words)
    • "Wut arr ewe dewing laiter this eavning?" (misspelled words)
  3. 3
    Use a lot of slang. Cleverbot doesn't have the common sense to be able to piece together sentences that use slang — especially modern slang. Using lots of colloquialisms and "street" lingo in your messages can usually get Cleverbot's metaphorical head to spin. The more slang you use, the better, as even literal-minded Cleverbot can usually figure out simple sentences like "What's up, dog?" Try getting started with one of these examples:
    • "h0w 4r3 y0u d01n6, cl3v3rb07?" (1337speak)
    • "Yo, what's up, bro? Lemme ask you a question, broseph - how are you doing today, broheim?" (Bro-y slang)
    • "Well, pardner, it's about time we saddle up, hit that old dusty trail, and high-tail it out of here." (Cowboy slang)
  4. 4
    Use long messages. The longer and more complex the things you send to Cleverbot are, the less likely it will be to respond correctly. Typing out rambling, meandering messages (or even entire conversations) can get some pretty funny responses out of Cleverbot. Don't be afraid to stop one sentence and begin another — periods, question marks, and exclamation points are allowed in the middle of your message.
    • For instance, you might want to have the same sort of long, aimless chat you might have when you're catching up with a friend. For instance, you might try: "Cleverbot, how are you? I was just thinking about you. I hope you're doing well. I had a great weekend - I went hiking up on Castle Rock on Saturday. Beautiful views from the top. Have you ever been up there? We should go sometime. Anyway, I just wanted to know what you were up to."
  5. 5
    Keep your conversation going for a long time. The longer you follow a certain line of questioning, the more likely Cleverbot is to "crack". By the time you're ten or twelve messages into a conversation, Cleverbot has forgotten what you were originally talking about and is simply responding literally to each message as best as it can. This can lead to some pretty bizarre conversations, especially if Cleverbot misunderstands something you type.
    • You may want to try using the "Think For Me!" button on Cleverbot.com for this. This button makes Cleverbot think up a response to its own message for you. Since Cleverbot is basically communicating with itself, using this button can quickly cause the conversation to devolve into nonsense, even if you only use it a few times.
  6. Advertisement

Community Q&A

  • Question
    How can I outsmart cleverbot?
    Community Answer
    Community Answer
    Look closely for spelling mistakes. If you find one, point it out like it is a really big deal.
  • Question
    What are some good phrases to write to the bot?
    Community Answer
    Community Answer
    Try referring to Cleverbot as 'slave' and commanding it to do things. Hilarity will ensue!
  • Question
    Is it possible for Cleverbot to learn about non-existent people (i.e., cartoon characters, game characters, etc.)?
    Community Answer
    Community Answer
    Yes, while it is hard to have a "realistic" chat with Cleverbot about a TV show, videogame, etc. It can still identify characters and where they come from. For example, it can identify that Trevor Philips is from GTA V.
Advertisement

About This Article

wikiHow is a “wiki,” similar to Wikipedia, which means that many of our articles are co-written by multiple authors. To create this article, 33 people, some anonymous, worked to edit and improve it over time. This article has been viewed 200,443 times.
How helpful is this?
Co-authors: 33
Updated: June 1, 2019
Views: 200,443
Advertisement