ChatGPT – What do SMEs need to know?

Opportunities and challenges of OpenAI's new chatbot.

The American startup OpenAI released their new AI chatbot ChatGPT at the end of 2022. Since then, anyone can chat online with the chatbot, ask it questions, give it commands or even have it write code. What changes for SMEs?

Background

At the end of 2022, the US company OpenAI released the chatbot ChatGPT and since then the system has been available to all internet users for free. ChatGPT is a chatbot, i.e. software that simulates a human conversation. On the website there is a simple input window, like in a search engine. If you register on the site (e.g. with a Google account) , you can type in any questions or give commands for free.
The special feature of ChatGPT is that the system generates texts in a few seconds, which at first glance can hardly be distinguished from those written by real people. The bot writes poems, essays, news texts, song lines and even program code. While the system is still under construction, it has trained how language works with about 500 billion words.

The speech robot developed by the company OpenAI is based on the GPT-3 system, which is a very comprehensive artificial intelligence (AI) model that has been enriched with almost all Internet sources, such as Wikipedia articles, digitized book collections, Twitter tweets and many other text sources. With this content, the AI learns to recognize the patterns and structures of language. If you now give the chatbot a word, it can predict with a very high probability what the next word should be based on statistical correlations. So far, we know this method from apps like WhatsApp, which often suggest the next possible word to be typed to their user. As soon as the algorithm has calculated the next word, it starts again from the beginning based on the previous series of words and adds more and more. The result is a complete sentence, or even an entire text, poem or programming code. But although the machine can recombine the content it has learned, it never generates anything entirely new. Nevertheless, this is enough to make us marvel.

Initial feedback from companies

The Institute for Financial Services Zug (IFZ) conducted a survey at the beginning of 2023 with companies that already use chatbots or voicebots. The majority of the companies surveyed are active in the financial environment and have been using bots for more than a year.

The majority (63%) of the employees surveyed did not first learn about ChatGPT and the opportunity to try out the new technology online through the survey, but rather beforehand through social networks or other media, such as television, radio or daily newspapers.

The topic seems to be not only on the news or social networks. In 19% of the organizations surveyed, an internal news release was sent out to all employees due to ChatGPT and the opportunities and risks it presents.

In 72% of the organizations, the topic of ChatGPT has triggered new discussions. These are often ethical questions or questions regarding data protection and copyright. But also questions concerning future use cases for chat and voicebots and possible applications of ChatGPT are discussed.

As things stand today, it can be assumed that ChatGPT will also have a short- and long-term impact on Swiss companies.
19% of the respondents state that their team has already received more attention and importance within the company since the reports on ChatGPT. Others report that internal issues regarding data privacy and ethics have become significantly more important in recent weeks. However, just under half of the survey respondents (44%) admit that they have not noticed any changes within the chat and voicebot teams or projects so far.

For the future, respondents definitely hope for significantly more attention for their chat and voicebot team or project. 31 percent of the employees surveyed also hope for a significant expansion of the current bot solution.
While ChatGPT has already triggered a number of issues and discussions within companies to date, half of the respondents do not expect most of the impact to occur until Q3 2023. 13% expect greater impact earlier (in Q2 2023) and another 13% expect impact later (in 2024).

In addition to wanting more attention and prioritized implementation of chat and voicebot projects, thanks to ChatGPT, respondents also see new dangers in using AI technology. Half of employees are critical of the fact that Swiss companies have no influence on OpenAI’s AI models and could therefore lose control. 38% have concerns about their own data protection policies.

The entire figures of the qualitative survey can be read here. (Survey in German by Institute for Financial Services Zug (IFZ))

What makes ChatGPT so special

The hype around ChatGPT is insane. Hardly any ordinary chatbot from a classic company has caused such a stir or reached so many users. The following selected features of ChatGPT explain the hype:

  • ChatGPT is able to handle context and can also refer to previous replies.
  • ChatGPT does not only write single sentences, but is able to write coherent texts, poems, blog posts and the like.
  • ChatGPT can even write program code.
  • The expression of ChatGPT is in most cases on a very high level. A difference to texts, which are written by humans, is hardly possible at first sight.

What companies need to know about ChatGPT

Companies thinking about using ChatGPT or GPT-3 (the AI model on which ChatGPT is based) should especially keep in mind the following tips:

  1. Currently, all of GPT’s artificial intelligence training models are stored on American servers. European companies can use the models, but they have no influence on them and cannot adapt them at this time.
  2. So far, ChatGPT’s responses are based on data from the Internet through 2021, so they are never current responses. The system is not able to give daily updated information at the present time.
  3. Even if a large number of ChatGPT’s texts are correct, it is not impossible that the system makes mistakes. However, since the system formulates each answer in a very pleasant way, it is hardly noticeable at first glance when the system gives out untruths.
  4. Thanks to ChatGPT and the associated hype in the media, the number of people who have already chatted with a chatbot has increased significantly. Customers are now more accustomed to the technology of bots and perhaps dare to use them in other industries.
  5. ChatGPT has not only led to greater adoption of chatbot technology, it has also raised expectations for chat conversations with an AI. ChatGPT is able to understand context and formulate human-like text. Previous chatbots or voicebots are usually unable to do this at all, or only to a very limited extent.
  6. Users will now expect these competencies more and more from other bots.
    So far, there are still many unanswered questions regarding data protection and the use of ChatGPT in the enterprise. It is therefore important to wait and see or to actively deal with the legal situation.

And now?

This post has shown that ChatGPT is not just a simple hype, but really brings some amazing features. However, it has also become apparent over the last few months that some aspects need to be considered when using ChatGPT. This is especially true with regard to data protection and the controllability of the system.

Companies, especially SMEs, should by no means close their eyes and refuse to use the new technology. On the contrary, it is advisable to start looking at the new technological possibilities today. For example, it would be possible to test the technology internally for “own purposes” first.

Nach Stationen in der Industrie und in Agenturen gehörte Sophie Hundertmark zu den ersten Masterstudentinnen in der Schweiz, die zu Chatbots forschen. Seit 2021 promoviert sie an der Universität Fribourg zum Einsatz von Chatbots im Banking und Insurance-Bereich. Dazu arbeitet sie als wissenschaftliche Mitarbeiterin am Institut für Finanzdienstleistungen Zug der Hochschule Luzern (IFZ). Sophie Hundertmark verfügt über langjährige Erfahrungen als selbstständige Beraterin für die strategische Begleitung sowie Umsetzung von Chatbot-Projekten. In diesem Zusammenhang sorgt sie für einen regelmässigen Austausch zwischen Akademie und Praxis zu allen Themen der AI getriebenen Conversational Automation. Anfang 2022 hat Sophie zudem den Verein Greenwishing Schweiz gegründet und engagiert sich seitdem aktiv gegen Greenwashing.

Comments are closed.

This website uses cookies to improve your experience. We'll assume you're ok with this, but you can opt-out if you wish. Accept Read More