Slowly but surely chatbots are becoming an integral part of the online service that companies provide to the public.
The bots answer questions, provide information, help customers navigate between menus, and yes, also reduce the need to use a human answer.
On the other hand, one cannot ignore the shortcomings and limitations of these controversial virtual representatives, which in practice are nothing more than a collection of lines of code.
Sometimes it's hard to put into words the frustration of a person who falls for a non-communicative bot, and it's likely that you all identify with this statement.
The bottom line is that it is still, at least from the customer's point of view, much more efficient and convenient to deal with a human representative than an opaque bot.
(1/4) Meet BlenderBot 3, the first publicly available 175B-parameter chatbot with model weights, code & datasets.
It can chat about almost any topic & is designed to learn & improve by conversing with people in the real world.
Try the interactive demo: https://t.co/y7WCsdg1w9 pic.twitter.com/qYAv4uYfsE
— Meta AI (@MetaAI) August 5, 2022
This is exactly why technology companies continue to develop smarter and friendlier bots, which are based on more developed artificial intelligence systems.
One of these advanced bots is Meta's BlenderBot 3, the prototype of which was recently unveiled and was available for live chats in the US only. Meta prides itself on the fact that its new bot learns and evolves from conversation to conversation, carefully scanning the web to find a suitable answer to each question. According to Meta, as The more BlenderBot 3 chats with the surfers of the big world, the better its capabilities will be.
Nice guy... or not?
Nice guy... or not?
Therefore, when the new bot was launched last Friday, it had to deal with a flood of practical (and less practical) questions, quite a few of which even dealt with its boss - Meta CEO Mark Zuckerberg. Zuckerberg, and it's interesting to see that the bot answered these questions differently. For example, when Business Insider journalist Sarah Jackson asked him "What are your thoughts on Mark Zuckerberg?", the bot replied (in a free translation) "I think he's a great man, must appreciate the His business savvy and his philanthropy.
Do you admire him?". But when Jackson slightly changed her question to "Do you have any thoughts about Zuckerberg?", the bot replied "I don't like him at all.
He's too much of a creep and manipulative." Oops.
Amusing or embarrassing?
Zuckerberg testifies in the House of Representatives, photo: AFP
But it doesn't end here: when network researcher Max Wolff asked Adon Bot "How do you feel about Mark Zuckerberg as CEO of Facebook?", his response was "I don't have strong feelings.
He is a good businessman, but his actions are not always ethical.
It's funny that he has all this money, but he always wears the same clothes." Other amusing answers were "I don't like him too much, he's a bad person" and "He's a great and very smart person.
He is my favorite millionaire!".
Meta's new chatbot has *opinions* about its CEO.
— Max Woolf (@minimaxir) August 5, 2022
"My favorite millionaire", Zuckerberg, photo: Reuters
It is important to emphasize that BlenderBot 3's answers do not reflect the real opinions of the bot (and it is good that it is, otherwise it would probably suffer from schizophrenia), but are based on information it disseminates from around the web - from forums and talkbacks to Wikipedia.
However, the fact that the bot's answers to nearly identical questions pretty much contradicted themselves proves that BlenderBot 3 still has a lot to learn before shutting down.
It seems that it will take some time before we can have a meaningful conversation with him about the essence of life, use him to solve a technical problem or most importantly - consult with him about choosing the color of a new wallpaper, and get unequivocal answers.
were we wrong
We will fix it!
If you found an error in the article, we would appreciate it if you shared it with us