Title goes here

About us
Total 39 companies invested
Total 79 companies invested

OpenAI and ChatGPT is a revolution, but the machines won’t take your job

Our analysts at Joint Journey view the emergence of new AIs, such as ChatGPT and DALL-E, as the arrival of a new means of production. Despite the common belief that advanced AI is intended to replace human labor, I don't think this is the case. Let's explore what the arrival of this new AI means for markets and venture investments.

In 2022, AI was democratized and became accessible to everyone, unlocking visual and text content generation, advanced search, and analytics. Seasoned machine learning engineers may not view this as a recent revolution, as they have been utilizing AI for several years. However, regular people have never had access to anything like this before.

What is happening now is a revolution in means of production, as there is now a low entry threshold, making AI available to everyone. Very soon, you will see everyone using AI for their everyday work and needs.

What is AI good (and bad) at?

Let's take a closer look at what AI is capable of and what its limitations are. Traditionally, people tend to view AI as a substitute for humans, perhaps influenced by impressive sci-fi narratives about the rise of machines or simply a cognitive bias. It is true that AI has successfully replaced some human positions, which may lead us to seek out more areas where AI can substitute for human tasks:

  • Can AI handle customer support inquiries on the 1st line?
  • Can it generate social media posts for me?
  • Can it assist with medical diagnosis?

And then we find the areas where AI fails to substitute humans. How should we approach this? It is misleading to assume that AI is simply not good enough, and we just need to improve it. This perspective is unproductive and overlooks the true potential of emerging AI, even if we ignore the ethical implications of wanting AI to replace more and more humans. So, what is the actual value that AI can provide?

Tame an AI and it won’t take your job

Let's take a closer look at how AI products interact with users and how happy users are with their experience.

When DALL-E creates an image for a marketer, it is always reviewed before being used. This is in accordance with the marketer’s expectations.

On the other end of the spectrum are chatbots in technical support. While they have made significant progress in recent years and can help clients in many cases, you might remember insisting on talking to a human operator no matter how hard the AI resists. There is a conflict where the business saves on support personnel, but users are not as content with it.

However, there are alternative ways for AI into tech support. An AI can monitor the communication between a support operator and a client (through chat or voice) and suggest a script in real-time. This is similar to DALL-E's images suggested to marketers’ review: the support operator is aware that it is an AI that may make errors, and verifies the recommendations. In this scenario, the AI is an assistant to the operator, and there is no conflict between the business and the operator about its use.

So, when AI fails to replace a human, it holds promise as an assistant.

Assistants, creativity and blank page syndrome

Think of the new AI as a junior manager on your team. A hard-working one, who can stay focused on many facts and aspects. But who sometimes makes very silly mistakes, because they’re junior. Or even think of an outsourcing agency that develops a mobile application for you. Would you trust your junior's work unsupervised? Would you leave an agency's operation on its own? Most likely not. Similarly, an AI needs your attention.

We use tools like ChatGPT or Midjourney to write or draw for us, but we tend to forget that this process isn't creation. It's a combination of analytics and synthesis, something the new AI excels at. Ask an AI to search for information and make a report or to suggest 50 ideas for a marketing campaign, and it will provide suggestions based on vast amounts of data. They are accurate often enough that it pays to use the AI with supervision. They might also be something you won’t expect to create yourself.

Interestingly, people tend to perceive AI's outputs - pictures, texts, and brainstorming ideas - as creative, but these outputs are actually derived from existing data. AI has access to the creativity of millions of people, allowing it to generate multiple ideas that you would deem as creative. What we take for creativity is actually the averaged creativity of many humans. AI doesn’t think outside the box, but its box is very big.

And yet, paradoxically, AI can help you get out of your own small human box when you're stuck staring at a blank page. I think the new AI’s place will be the "supervised assistance." It can handle the grunt work, by it also can kick-start your creativity when you're exhausted. It will provide you with good practices and the best ideas, saving you from endless Googling and eventual collapse from the amount of data to process.

Applications for business

Here are some of my expectations for how the new supervised AI will augment businesses in the near future:

Business Processes
  • Technical support: generating scripts, scoring and prioritizing customer chat messages, emails, and calls, and analyzing tone-of-voice.
  • Assisting product and project managers in writing requirements for business processes and products.
  • Scoring leads for sales and marketing.
  • Translation and localization.
  • Brainstorming for product development, marketing, and sales.
  • Copywriting for content marketing and social media marketing.
  • Generating images for ad campaigns, social media, and articles.
  • Writing code.

User Experience Personalization
  • Building career paths in education.
  • Writing dialogues for games.

What’s next?

  • As a revolution in means of production, the new AI has evoked an avalanche of new tools. In a few years, AI-based tools will enter our everyday lives and influence everything we do, just as Google search, smartphones, YouTube, and social media have done in past two decades.

  • New toolkits for dedicated industries will emerge, requiring training. Just as the industrial revolution created a class of engineers, the new revolution will create a highly demanded class of AI engineers.

  • If you want to develop a new AI-based tool, you can start with identifying the part of your job that is mundane or repetitive. This is usually the thing you would want to outsource to an AI.

  • ChatGPT is a general-purpose AI, and it may not be suitable for narrow applications. New AIs specialized in particular areas will be developed soon and will perform better in their respective fields.

  • One of the most important questions an investor can ask is how a business plans to use generative AI, regardless of its industry. It's not about man versus machine, but rather about the competition between a person without AI skills and a person without them. And we all know who will come out ahead.