Interview with Gerd Leonhard
The generative AI tool ChatGPT has gained exponential interest and concern in equal measure since it launched in November 2022. Reaching 1 million users in just five days, ChatGPT is expected to bring its pioneer OpenAI $200 million by the end of the year. Before businesses jump on the ChatGPT bandwagon, it’s important to understand the opportunities and risks the tool presents, as well as its ethical considerations and limitations.
In this exclusive interview, futurist and thought leader Gerd Leonhard shares his thoughts on ChatGPT and gives sound advice to business leaders on how to leverage and manage this emerging iteration of AI.
What is ChatGPT, and what are its limitations?
It’s a language learning model, an AI software that looks for patterns and clues on how to answer a question by looking at millions of statements that are related to the question. Of course, language models and AI have been around for 20 years with deep learning and machine learning. But the fact is that OpenAI decided to release ChatGPT to the public, creating a moon landing moment. ChatGPT is a bit like Sputnik, and the people wanting to go to the moon are Bing and Google …trying to integrate it. So, it’s a big deal for many reasons. I think it has great potential but also grave concerns like every big technological leap.
However, ChatGPT will not be like the Metaverse, cryptocurrency, or blockchain. It’s a real game-changing moment and something we should be looking at across the board.
Will ChatGPT replace existing search engines?
Search engines have been integrating ChatGPT gradually. For example, You.com and Bing, and Google soon — in the sense of having an optional chat bar. But we have to be careful of the relevance of those answers and fact check – which ChatGPT can’t do in a meaningful way currently. AI is a tool that we have to remain skeptical of. To paraphrase Kevin Kelly, “humans are for questions, machines are for answers.” I think that is so true. In this case, we should not be overestimating the truth and the validity of the answer.
I think the key problem is that it’s so tempting. ChatGPT may turn into a huge laziness-machine if we are not careful. With OpenAI’s recent partnership with Microsoft, we can expect fast development but it’s not going to replace search engines. Something to keep in mind is that humans make decisions by looking at a decision tree, images, and feelings.
AI knows nothing about real life, it only knows data life. So, we should remain careful as to how much we rely on AI.
What opportunities does ChatGPT present to businesses?
At this point, it’s at a very early stage at the commercial level. It's not scalable. It's not real-time. It does not have ethical guardrails. I think it's far away from large-scale commercial use.
There’s an old saying that goes, “one machine can replace 100 ordinary workers, but no machine can replace one extraordinary worker.” A person using generative AI will beat a person without it. The tool alone won’t beat anybody.
It will be the same for businesses. If you don't use these tools to become faster, smarter, and more efficient, you’re going to get left behind. I've said this for a long time, but whatever can be automated, digitized, robotized, virtualized, and chatbot-ized; will be. That has a significant impact on our structure, companies, profit margins, and so on. But that doesn't mean we will be out of work. Instead, AI frees up my time to do more meaningful work. So far, ChatGPT presents opportunities to the finance, customer service, and airline industries.
Do you think jobs will be lost to ChatGPT?
Technology needs to be used in a wise way. Look at call centers for example, I think 90% of those jobs will be automated but there will still be a need for supervisors. But yes, it will have an impact on routine jobs.
“If you work like a robot, a robot will take your job. If you learn like a robot, you'll never have a job, or you end up working for the robots”. Think of Maslow’s Hierarchy of Needs, the same is true for jobs. The lowest level is data, information, and simple binary law knowledge. Machines have an unlimited repository of knowledge, and this will be clear by 2030. We are moving to the next level of work which is about tacit knowledge, quiet knowledge, understanding, wisdom, and purpose. There are plenty of jobs there. Also, I think if you’re going to save 50% of operating money using AI, that money needs to be put back to create possibilities of re-education and re-skilling.
Can ChatGPT measure specific outcomes in terms of customer satisfaction?
I think it could do that well. Imagine feeding all your data into an internal version of ChatGPT. You could ask intelligent questions like, “Do customers really like this product?” It could yield some astonishing results. Of course, data security and safety are issues here. In addition, a lot of data that humans use is not data that machines get. So, I think we need to use it for trivial work. For example, travel agencies can utilize ChatGPT to figure out where their customers most likely want to go. From there, they can create a web page with that offering, alter codes on existing web pages, and more.
How can ChatGPT be used to generate creative and engaging content for marketing and sales purposes?
There’s a great article by The Atlantic about how prompting AI is becoming a skill. Prompting means giving the right commands but if you drill deeper and ask more complex questions, you will get more complex answers. So, you need to be good at prompting. One thing businesses can try immediately is to prompt their own products. You can ask how customers are receiving the product, what they don’t like about the product and more.
However, keep in mind that it’s not real-time as ChatGPT’s database stops in 2021. That is one of the great confusions about ChatGPT and generative AI in general. I don't think it's possible to make it real-time. Because real-time is an infinite universe with data doubling every two weeks. That is where search engines have an advantage over generative AI.
What is your advice to business leaders who want to implement ChatGPT in their businesses?
Firstly, experiment and try everything but be careful at the same time. For example, many companies thought that social media marketing could replace all marketing activities and save them money. Turns out that wasn’t true. Social media is expensive. So, you're not going to save money because of this tool, you’re probably going to save some time. Appoint someone to oversee this and go for low-hanging fruits. For instance, explore bots for customer service and microsites.
Secondly, we should embrace technology but not become technology.
We should establish clear borders as to what it's good for what it's not. You don't want to risk losing the trust of your customers. Too much of a good thing can be bad, and that is true with technology. Too much technology can really make a mess in your organization and customer communication.
*The interview answers have been edited for length and clarity.