Don’t worry, AI can be managed by humans

We will all be responsible for deciding how much of creative work is ultimately done by AI and how much is done by humans.

We can manage AI.

In the face of technological change, creativity is often held up as a uniquely human quality, less vulnerable to the forces of technological disruption. But today, generative artificial intelligence applications like ChatGPT and Midjourney are threatening to up-end this special status and significantly alter creative work, both independent and salaried. These new generative AI models learn from huge data sets and user feedback, and can produce new content in the form of text, images and audio or a combination of the three. As such, jobs focused on delivering content – writing, creating images, coding and other roles that have historically required an intensity of knowledge and information – now seem likely to be uniquely affected by generative AI.

We propose three possible – but not mutually exclusive – scenarios for how this development might unfold and offer recommendations for what companies should do to prepare for this brave new world.

AN EXPLOSION OF AI-ASSISTED INNOVATION

Not quite two years ago, GitHub introduced GitHub co-pilot, an AI “pair programmer” that helps people write code. More recently, designers, filmmakers and advertising executives have started using image generators such as DALL-E 2. These tools don’t require users to be very tech savvy. Pretty much anyone can make use of them.

In this scenario, rather than putting many creators out of work, AI will support humans to do the work they already perform, simply allowing them to do it with greater speed and efficiency. This would lead to a rise in productivity, as reliance on generative AI tools that use natural language reduces the time and effort required to come up with new ideas or pieces of text. Of course, humans will still have to devote time to correct and edit the generated information, but, overall, creative projects should progress more quickly.

With reduced barriers to entry, we can expect more people to engage in creative work. GitHub’s co-pilot doesn’t replace the human writing code, but it does make coding easier for novices. If more people learn “prompt engineering” – the skill of asking the machine the right questions – AI will be able to produce very relevant and meaningful content that humans will only need to edit before they can put it to use.

MACHINES MONOPOLISE CREATIVITY

A second possible scenario is that unfair algorithmic competition and inadequate oversight will lead to the crowding out of authentic human creativity. Here, human creators are drowned out by a tsunami of algorithmically-generated content. If that were to happen, an important question to address would be: How will we generate new ideas?

A nascent version of this scenario might already be happening. For example, recent lawsuits against prominent generative AI platforms allege copyright infringement on a massive scale. What makes this issue even more fraught is that intellectual property laws have not caught up with the technological progress that has been made in the field of AI research.

In this scenario, generative AI significantly changes the incentive structure for creators, and raises risks for businesses and society. If cheaply made generative AI undercuts authentic human content, there’s a real risk that innovation will slow down over time as humans create less. Creators already face intense competition for human attention spans, and this kind of competition will only grow if there is unlimited content on demand. Extreme content abundance, far beyond what we’ve seen with any digital disruption to date, will inundate us with noise. Yet even in this relative dystopia, we will still need humans to curate the existing content in this ecosystem.

‘HUMAN-MADE’ COMMANDS A PREMIUM

The third potential scenario that could develop is one where the “techlash” resumes with a focus against algorithmically-generated content. One plausible effect of being inundated with synthetic creative outputs is that people will begin to value authentic creativity more and may be willing to pay a premium for it. While generative models demonstrate remarkable capabilities, they suffer from problems with accuracy, frequently producing text that sounds legitimate but is riddled with errors. For obvious reasons, humans might demand greater accuracy from their content providers, and therefore may start to rely more on trusted human sources rather than machine-generated information.

In this scenario, humans maintain a competitive advantage against algorithmic competition. The uniqueness of human creativity will become important leverage. Culture changes much more quickly than generative algorithms can be trained, so humans maintain a dynamism that algorithms cannot compete against. In fact, it is likely that humans will retain the ability to make significant leaps of creativity, even if algorithmic capabilities improve incrementally.

In this scenario, content moderation needs are likely to explode as information platforms are overwhelmed with false or misleading content, and therefore require human intervention and carefully designed governance frameworks to counter it.

HOW TO PREPARE FOR GENERATIVE AI

Below, we provide three recommendations that workers should consider as they adopt generative AI to create business value and profit in today’s creative industries.

Prepare for disruption, and not only to your job. Generative AI could be the biggest change in the cost structure of information production since the creation of the printing press in 1439. The centuries that followed it featured rapid innovation, sociopolitical volatility and economic disruption as the cost of acquiring knowledge and information fell precipitously. We are in the very early stages of the generative AI revolution. We expect the near future to be more volatile than the recent past.

Invest in your ontology. Codifying, digitising and structuring the knowledge you create will be a critical value driver in the decades to come. Generative AI and large language models enable knowledge and skills to transmit more easily across teams and business units, accelerating learning and innovation.

Get comfortable talking to AI. As AI becomes a partner in intellectual endeavours, it will increasingly augment the effectiveness and creativity of our human intelligence. Knowledge workers will therefore need to learn how to best prompt the machine with instructions to perform their work.

Get started today experimenting with generative AI tools to develop skills in prompt engineering; a prerequisite skill for creative workers in the decades to come.

David De Cremer is the provost’s chair and professor in management and organisations at NUS Business School, National University of Singapore. Nicola Morini Bianzino is EY’s global chief technology officer. Ben Falk is a director in EY’s chief technology office.

Copyright 2023 Harvard Business Review/ Distributed by NYTimes Licensing