If you are old enough to remember, you may recall Microsoft’s first attempts at making a virtual assistant. His name was Microsoft Clippit, commonly known as “Clippy.” He was seen as the answer to all of Microsoft’s problems at the time of its creation in 1997. Many people were still getting used to the idea of computers, viewing them as complicated, intimidating machines rather than convenient user-friendly tools. Microsoft attempted to change people’s perceptions with an adorable assistant. But unfortunately, it wasn’t the change Microsoft was hoping for. Rather than seeing Clippy as a warm, welcoming assistant and a new addition to the Windows experience, many people found him to be annoying, distracting, and further complicated everything. When trying to perform basic tasks such as writing a letter in Word, your “assistant” would get in the way. This was often seen when he asks bubble-headed questions or provided irrelevant advice that would ultimately hinder productivity. Microsoft’s mission to facilitate the use of its products failed, even in some extreme cases turning off Windows entirely for people. When looking at Clippy in retrospect, it is quite obvious that it was destined to fail from the beginning.
The Concept & Idea of Clippy.
On January 24th, 1984 the very first Macintosh was released. A Personal Computer would be much more personal. Although by no means the first of its kind, the Macintosh took technologies of its kind and made them very mainstream to a lesser extent with Text-to-Speech, and to a much greater extent, a Graphical User Interface. You could perform what was considered the most sophisticated of tasks at the time with just the click of a button. You no longer exclusively needed a command line. By adding a voice and a unique look to this computer, the Macintosh now had a personality. It further taught the industry that computers do not need to be utilitarian calculators. They can be useful tools. The Macintosh was of course not immune to its problems, but computers moving forward would take great inspiration from the concept that the Macintosh introduced. Obviously, this would also mean competing vendors. Computers were now easy to use.
A year after the Macintosh was released, a new Operating System would be released. One which Steve Jobs alleged was completely stolen from Apple. This would be Microsoft Windows. The fact that this Operating System was compatible with a variety of machines, made it accessible rather than a $2,500 Macintosh. Naturally, Microsoft’s influence in software grew. By the 1990s, they would be known to almost everyone. Especially with their release of Windows 95. With practically the entire computing industry in their hands, Microsoft wanted to revolutionize the art of personal computing even further. In 1996, research was conducted at Stanford University which was put together by communications professors Clifford Nass and Byron Reeves. Those who have been reading for a while may recognize this research from an article that I wrote last year on BonziBuddy. There is some overlap between BonziBuddy and Clippy. In their findings, Nass and Reeves concluded that people treat computers the same way they treat television, advertisements, and media outlets. Overall, people treat computers like other humans. Despite these being inanimate objects, people would develop relationships with them, particularly with voice-based technologies. Studies show that people feel more welcomed and safer around computers that use female voices rather than male voices. “We’re all human.” As some say. When we see something that is more complicated, robotic, or the exact opposite of what we are, it can be intimidating and prevent people from wanting to interact with it. This theory is known as The Media Equation.
The Release of Microsoft Bob.
Microsoft’s new mission was to use this theory as its blueprint for developing products. Luckily, Microsoft wasn’t completely oblivious to the possibility of a theory like this surfacing. You could say Microsoft predicted this theory. All in the wrong ways. One year prior to The Media Equation being published, Microsoft Bob was released. This was a program designed for Windows 3.1, the version of Windows before 95. This made the GUI experience much more personal and literal. While the Macintosh was revolutionary making complex code appear as symbols or objects people use, Microsoft had a different approach. This was done by making the computer appear as an entire room. For example, if you needed to set an alarm, you would click on the clock on the wall. If you have questions, your assistant Rover would help. This could not be considered a computer, it was your second home. Just in the digital world. While this idea may sound revolutionary, it was a complete failure. The main takeaway from this is that Microsoft misunderstood its audience. While this serves the objective of being simple, Microsoft seemed to treat its consumers as if they lacked intelligence. It would be difficult to see this taken seriously in a workplace environment. Especially in the ‘90s.
In regards to The Media Equation, Microsoft unintentionally followed two of its main points. Those are the fact that people will use things the way they are represented, and people will use things that are simple. While the Macintosh was familiar, it still presented itself as a computer. There was a balance between the two extremes. Microsoft Bob however failed to do so. By leaving no room for interpretation, Microsoft effectively insulted the intelligence of its base. To put it simply, Microsoft misunderstood the first point and took the second one too far.
Microsoft Clippy.
With Windows 95’s successful release, Microsoft realized that there was no need to overanalyze everything. However, Windows could be more straightforward. As a result, The Media Equation was taken into account where Windows should have a personality. Microsoft decided to keep the assistant aspect of the product known as Microsoft Actor with minor changes. Instead of having this assistant on the entire Operating System, Microsoft Actor was seen on Office with “Office Assistant.” Considering this assistant would only be seen in Microsoft Office, it would make sense to have it office-themed. Kevan Atteberry, known for his illustration for Microsoft Bob designed a character. After Atteberry concocted some sketches, he created his new character. This would be known as Microsoft Clippy. With the release of Office 97, Clippy was now part of Microsoft’s list of characters. Ready to answer all your questions.
Unfortunately, this would not be the case. As Clippy would offer unnecessary help. Such as asking if you need help writing a letter. Those on a computer at the time knew how to write a letter. Most users in the ‘90s weren’t children. They are adults often in work environments. It seemed that Microsoft was repeating one of the many mistakes they made with Microsoft Bob, treating their user base as if they lacked intelligence. These users do not want to be talked down to. Microsoft did not understand their audience. Although Clippy was the one that succeeded the most in focus groups and therefore was made the default assistant, Microsoft still willingly ignored one problem that seemed small but would ultimately make a massive difference. Referring back to the study mentioned earlier about computers with male and female voices in which users had a more positive experience with the female computer, Clippy had the opposite issue. It seems that Clippy was too masculine. Women in these focus groups felt uncomfortable by him as if he was staring at them in a flirtatious way. Men running the focus group obviously dismissed the feedback completely. While Clippy was the default assistant, there were more. Other assistants included Hover Bot, Shakespeare, and Einstein. But they are still masculine characters. If we were strictly following the results of this study, it makes sense why user experience was more negative than if these characters were more diverse. But this was just one out of many problems.
The Problems With Clippy.
In addition to the overly cute characters that appealed to children, the design of the technology itself was also annoying. Clippy would often get in the way distracting what could otherwise be very focused people. While you could always turn Clippy off in your Office Settings, most people didn’t know how to do that. Clippy was also around at the very worst time. Lastly, it didn’t help that Clippy was more or less a personified version of Office’s existing Help Menu. Clippy was pre-programmed and felt pre-programmed, repeatedly asking the same questions and not taking users’ preferences into account. Clippy wasn’t a personal assistant. Clippy was very mechanical. It was actually counterintuitive in its objective to fulfill the propositions postulated by The Media Equation. When taking everything into account, however, is it really Microsoft’s fault? Was Clippy too ambitious for his time? Keep in mind that this was before sophisticated machine-learning-based AI or accessible virtual assistants such as Siri.
In many respects, Clippy came out at the wrong time. What it wanted to do was limited by the technology of the ’90s. With that said, you could also take poor execution into account. Considering BonziBuddy used the web for things such as looking up answers that users submitted gave it somewhat of a personality and made it functional. Even if its motives were different. Although the internet was in its infancy in 1997, Microsoft could have very well updated Clippy with these web searching capabilities as the years went by. But even then, the damage was already done.
The Possible Rise of Clippy.
By the time Office XP came around, the message was clear. Clippy was turned off by default. You could turn him back on if you wanted to. Clippy also had an improvement. But Microsoft already knew that they have made a mistake with Clippy, putting him at the butt of many jokes and adding references within their programs. The actual functionality of Clippy would be removed in 2007. However, there is possible talk of bringing him back. There are currently rumors that Clippy could return in the way that it was supposed to. Especially with Microsoft’s investment into OpenAI. Clippy could very well become a personification of Chat GPT, effectively changing it from the worst AI tool to the best one in history, even outbidding Cortana. If that is true, that would be quite the redemption.
Although Clippy failed to make a connection with its audience in a technologically useful sense, it certainly did in an emotional sense in the long term. Many of those who hated Clippy now love him. In addition to his obnoxious aura just becoming pure meme material, another reason is nostalgia. There are now grown adults whose earliest memories are of Clippy. For many people, he is the embodiment of the 1990s and early 2000s. Clippy takes them back to those much simpler times, their childhood. Clippy still lives in our everyday lives. We could attribute the AI Software that we use today to the failure of Clippy. It was a valuable lesson that taught the technology industry what not to do when creating virtual assistants. That is absolutely not a failure.
