The AI industry is changing. Generative AI, as well as other foundation models, are bringing assistive technology to a whole new level. They also reduce application development times and bring powerful capabilities to users who are not technically minded.

In keeping with the topic of this article, we asked ChatGPT – a generative AI model that can generate original content by responding to a prompt from the user – how it would set the context. The system was given the following prompt:

In the style of Kobe Digital, write the first paragraph for an article on how transformative generative artificial intelligence can be for businesses.

ChatGPT’s response:

The rise of generative artificial intelligence has the potential to change business forever. This technology can revolutionize industries by allowing them to create original content using existing data. Generative AI can increase productivity and efficiency by automating tasks previously performed by humans. It also reduces costs and opens up new growth opportunities. Businesses that can effectively leverage this technology will likely gain a competitive advantage.

Not perfect, but still incredibly impressive.

We and other executives with whom we spoke recently were able to experience the power of generative artificial intelligence by simply entering a single prompt into ChatGPT. This was developed by OpenAI. More than one million users visited the platform in the first five days after its launch to test it out. OpenAI’s server can barely keep pace with the demand. Users are regularly prompted to return later when there is more capacity.

ChatGPT, GitHub Copilot and the underlying AI models (Stable Diffusion DALL*E2, GPT-3 to name a couple) are bringing technology to realms that were once reserved for humans. Computers can exhibit creativity with generative AI. The computers can create original content to answer queries, using data they have ingested as well as interactions with users. They can create blogs, design package designs, code, or theorize about the cause of a production mistake.

The foundation models are large-scale deep-learning models that have been trained using massive, broad, and unstructured datasets (such as images and text) covering many topics. The models can be adapted to a variety of tasks, and little fine-tuning is required. GPT-3.5 (the foundation model that underlies ChatGPT) has been used for translating text and scientists have used an older version of GPT to create new protein sequences. This makes the power of these abilities accessible to everyone, even developers without specialized machine-learning skills or, in some cases people with no technical backgrounds. The use of foundation models can reduce the time required to develop new AI apps.

AI will be more exciting in 2023 with the advent of Generative AI. As with any new technology, business leaders must be aware of the ethical and practical implications.

Pushing into the human realm

We wrote an article more than a decade back in which we divided economic activity into three buckets: production, transactions, and interactions, and examined how technology had impacted each. Machines and factory technology transformed production during the Industrial Revolution, more than 100 years ago. AI has also increased efficiencies on the manufacturing floor. Over the past 100 years, transactions have gone through many iterations of technology, most recently digitalization and automation.

Up until recently, the most mature technology interventions were in interaction labor such as customer service. The emergence of generative AI will change this by tackling interaction work in a manner that closely resembles human behavior, and in some cases is imperceptible. These tools are not designed to function without human input or intervention. They are often most effective when combined with humans. This enhances their abilities and allows them to do work faster.

The technology of Generative AI also pushes the boundaries of creativity, a domain that was thought to be exclusive to humans. The technology uses its inputs (the information it has ingested, and a prompt from the user) and experiences to create new content. Although dinner table arguments will continue to rage on whether or not this is true creativity, the majority of people would agree that these tools are likely to inspire more creativity in the world because they provide humans with starter ideas.

Uses for business are numerous

We’re still in the early stages of scaling these models, but the first applications are starting to appear across functions.

• Marketing – Creating personalized marketing, social media, and technical sales content, including text, images, and videos; creating assistants aligned with specific businesses such as retail
• Operations–generating task lists for efficient execution of a given activity
• IT/engineering–writing, documenting, and reviewing code
• Risk and Legal – Answering complex questions and pulling together vast amounts of legal documents, drafting and reviewing reports, and more
• R&D — accelerating drug discovery by better-understanding diseases and discovering chemical structures

Excitement is a good thing, but caution must be exercised

It’s easy to think that generative AI is a technology ready for use, but this is not true. Executives must proceed with caution due to its nascency. The technology is still in its infancy, and many practical and ethical questions remain unresolved. Here are a few examples:

• Just like humans, generative AI is not perfect. For example, ChatGPT can “hallucinate,” which means it generates completely inaccurate information as a response to user questions and does not have a built-in way to alert the user to this or to challenge the results. We have seen instances where the tool has been asked to create a bio, and the result was incorrect. For instance, it listed the wrong institution of education.
• The filters aren’t effective enough yet to stop inappropriate content. When users of an application that creates avatars based on a photo of a person entered the appropriate pictures of themselves, they received avatars that showed them as nude.
• There is still a need to address systemic biases These systems use massive amounts of data, which may include unwanted biased.
• The values and norms of individual companies are not reflected. To incorporate culture and values into technology, some companies will require technical expertise.
• Intellectual-property questions are up for debate. When a generative AI model brings forward a new product design or idea based on a user prompt, who can lay claim to it? What happens if it plagiarizes an idea or product based on the training data it has?

Executives: Initial steps

Companies that are considering generative artificial intelligence will need to identify areas of their business in which the technology can have an immediate impact. They should also implement a monitoring system, as the technology is expected to change quickly. It is a wise move to put together a team of data scientists, lawyers, and business leaders to answer basic questions.

• What is the impact of technology on our business, our industry, or our value chain?
• What are our policies? Are we waiting to see if technology changes, investing in new pilots, or trying to create a brand-new business? Should the approach be different across the various areas of business?
• What are the criteria we use to select the use cases that we target, given the limitations of our models?
• How can we build an ecosystem that is effective in terms of bringing together partners, communities, and platforms?
• Which legal and community standards must these models adhere to maintain the trust of our stakeholders?

It’s important to encourage innovation in the workplace. This can be done by setting up a sandboxed environment for testing, which is readily available through the cloud.

The innovations that generative AI could ignite for businesses of all sizes and levels of technological proficiency are truly exciting. However, executives will want to remain acutely aware of the risks that exist at this early stage of the technology’s development.