A Brief Biography of Rollo Carpenter: The Mind Behind Clever Bot

Last Updated on November 11, 2020 by Sean B

The field of Artificial Intelligence is filled with many instances of brilliant minds who came up with programs and bots, Rollo Carpenter is one of these brilliant men. He created chatbots and other work which revolutionized the field and helped others within the field of AI achieve milestones they would not have been able to reach were they not standing on Mr. Carpenter’s shoulders..

An example of these instances comes from the bots which did exceptionally well during the Turing Test, first proposed by brilliant mathematician and logician Alan Turing in 1950. The Turing Test is a test where judges hold a conversation with a human and a machine, and the job of the machine is to convince the judge that he is talking to a human.

In this article, we’ll take a look at the chatbots that Rollo Carpenter developed, and we’ll discuss some other aspects of his life.

Early Life and Career

Rollo Carpenter was born in 1965 in Britain. Rollo Carpenter had started programming at a very young age and was exceptionally gifted at handling databases. Ideas for creating chatbots clicked in his mind when he was working with programming languages.

Rollo Carpenter has been active in the field of AI throughout his career. Rollo Carpenter’s career started when he co-founded Bizfinity, Inc in 1997. Bizfinity is a software company that provides a business synchronization platform for small and growing businesses.

Before coming up with ideas for Jabber wacky and Clever Bot, Rollo Carpenter consistently experimented with AI and machine learning algorithms and developed many ideas for his chatbots.

Rollo carpenter always aspired to create a Chatbot that can learn from its interactions with humans and reply to humans based on those past experiences. Rollo Carpenter wanted a Chatbot that can be fun to talk to and does not feel like a machine completely devoid of emotions and compassion.

Development of Jabberwacky

Rollo Carpenter was working on Jabberwacky, sometimes called Jabber Wacky, as early as 1986, way before he participated in the annual Loebner Prize. Jabber Wacky aimed to stimulate human conversations in fun. Engaging, interactive, and humorous manner!

The strategy that came into his mind for creating Jabber wacky was to create a feedback system for the Chatbot that can help the bot learn from humans. His goal was to create a Chatbot that first learns from experiences and then uses that experience to talk to humans.

Jabberwacky is one of the earliest attempts to create AI programs that can engage with humans interactively and do not feel like emotionless machines with a dull environment around them. The goal of the Jabber wacky was to pass the Turing test and simulate conversations.

Jabberwacky was not meant to be used in computer support systems or corporate purposes; rather, it was specifically meant for entertainment. However, after the success of the Jabber Wacky chatbot, it did inspire many other Chatbots to be integrated into other sectors such as business, education, information, and finances.

The Chabot’s final goal is to evolve from a text-based system towards a voice-operated program that can learn from vocal interactions and give vocal responses accordingly. Rollo Carpenter aspires to create AI machines that can live alongside humans.

Jabberwacky eventually became Cleverbot.

Evolution of Jabberwacky Into Cleverbot

The Cleverbot is the updated and modified version of Jabberwacky. When Cleverbot went online in 1997, the bot was able to receive around 20,000 entries and carried out unique conversations with every individual rather than repeating the same pattern with every individual.

The quality that distinguishes Cleverbot from usual assistants such as Google Assistant is that interaction becomes predictable after a certain time as the pattern becomes apparent. However, the clever bot is capable of handling unique conversations with users and becomes more efficient with experience.

Further Developments In Cleverbot

The cleverbot learns from responses and effectively learns from the data collected by the users that interact with it from all over the world. In 2003, Cleverbot’s interactions crossed more than a million, and the figure incremented to 150 million interactions by 2019.

The clever bot is now also available as an IOS, Android, and Windows Phone App. Clever bot’s responses are not pre-programmed and are rather spontaneous, and the clever bot uses its experience with past conversations to generate appropriate responses.

Users can talk to the clever bot by heading to the site and typing in the description box below the Clever bot’s logo. The system looks for appropriate keywords in the user’s message and comes up with the most relevant response to provide to the user.

The clever bot is capable of handling conversations with 10,000 to 40,000 people at once. The clever bot is constantly growing in data size, with a rate of 400 to 7 million interactions per second. The clever bot has been consistently updated and improved behind the scenes.

In 2014, Clever bot was also upgraded to GPU (Global Positioning Unit). The clever bot uses fuzzy logic to converse with users rather than a fixed way.

Clever Bot’s Trip to India

On 23rd September 2011, a formal Turing Test was conducted in India by IIT Guwahati. IIT (Indian Institutes of Technology) are research institutes in India. Thirty volunteers conducted the test, and the conversation was typed four-minute.

Rollo Carpenter’s Clever bot also participated in this event alongside a human; while half of the people talked to the clever bot, the others talked to the human. The responses from both of the participants were shown on a large screen for the audience.

Clever bot performed exceptionally well. 1,334 votes were cast, that is way more than any previous Turing test. The clever bot was able to score 59.3% in the Turing test and was able to convince judges that they are talking to a human even though the winning requirement is merely 30%.

The human, on the other hand, scored 63.3% on the test. It’s remarkable how Cleverbot was able to convince so many people that they are talking to a human and not an AI program. This moment is a milestone for artificial intelligence.

Rollo Carpenter holding one of his two Loebner Prize Medals.


Entry in The Loebner Prize

The Loebner Prize is an annual competition that is based around the Turing test. It was first launched in 1990 by Hugh Loebner. The Loebner prize competition awards Gold, Silver, and bronze prices to the participants.

Jabber wacky participated in the Loebner Prize numerous times and won the prize two times, first as George and then the second time as Joan. With each entry, jabber wacky came back better than before and was able to improve its performance greatly.

Jabberwacky the 2003 Loebner Prize Entry

In September 2003, Rollo Carpenter’s AI program, Jabber wacky, participated for the first time in the Loebner Prize. Jabber wacky was awarded third prize in the competition and was beaten by Jabberwock, a German-based AI program that was created by Juergen Pirner.

The experience of jabber wacky was tremendous in terms of how it helped the Chatbot in improving itself. Jabber wacky was slowly getting better at conversations, as evident by its performance in each entry in the Loebner Prize competition.

Jabberwacky the 2004 Loebner Prize Entry

In September 2004, Rollo Carpenter’s AI program participated for the second time in the Loebner Prize. This time, Jabberwacky performed better than before and was awarded the second prize in the competition.

Jabber wacky was beaten by Chatbot ALICE (Artificial Linguistic Internet Computer Entity). Jabber wacky had improved by leaps and bounds and was learning new capabilities with time and experience.

George the 2005 Loebner Prize Entry

In September 2005, Rollo Carpenter’s AI program George who was an avatar or character created around Jabberwacky, participated in the Loebner Prize competition. This time Jabber wacky came out on top and won the Loebner Prize.

This was a defining moment for Jabber wacky as the trajectory of jabberwacky’s performance was going upwards as time went on. Just like Rollo Carpenter had versioned, his Chatbot was learning from experience, and of course, the adjustments made into him were also the reason.

Joan the 2006 Loebner Prize Entry

In September 2006, Rollo Carpenter’s AI program Joan, a character of Jabber wacky, participated in the Loebner Prize. Jabber wacky yet again won the Loebner Prize. By this time, jabber wacky had reached enormous potential, as evidenced by its winning streak.

A screen grab from a YouTube video featuring Rollo Carpenter.

Career after Gaining Prominence

After into prominence by his performance in Loebner prize competitions, in October 2005, Rollo Carpenter started working for Icogno ltd as its managing director. In 2008, Rollo Carpenter developed Cleverbot, a new and improved version of Jabberwacky.

After that, the Rollo carpenter worked for Existor ltd as its Managing Director. Existor is an AI platform with many famous Chatbots such as Eviebot, Biobot, Pewdiebot, and Chimbot. The platform also allows users to create their own AI programs.

Rollo Carpenter worked on AI programs related to entertainment, communication, companionship, and education. Rollo Carpenter worked as managing for more than one company.

Rollo Carpenter also worked as a CTO of a business software startup in Silicon Valley. In 2010, Rollo Carpenter won the British Computer Society’s Machine Intelligence Competition. Rollo Carpenter’s AI program Cleverbot is now considered one of the greatest chatterbots out there.

Rollo Carpenter holds the view that we still have a long way to go in terms of what AI can achieve, and he continues to work on AI programs that can break boundaries in the field of Artificial Intelligence.

In Closing

Rollo Carpenter is one of the most important figures in the development of AI technology, particularly as it applies to chatbots. He deserves to be highlighted for his two wins in the Loebner Prize competition with George and Joan, both avatars of Jabberwacky. But he should also be highlighted for the impact his work had on others. Jabberwacky is one of the reasons chatbots are so widely used today, and Cleverbot (particularly Evie) is in my opinion one of the more pleasant chatbots to talk to.

We hope that the blog was informative; feel free to leave your thoughts in the comments. 

Website | + posts
Share this:

Leave a Comment