EY position paper: AI-generated content in transition – between progress and fatigue

Read in German, read in French

• AI fatigue identifiable: While AI-generated content is shaping digital communication, people are increasingly looking for authentic offers. Traditional media formats, podcasts and interviews are experiencing a renaissance.
• Artificial intelligence is improving rapidly, yet transformative use in the corporate environment often fails due to the quality of corporate data, complex decision-making processes and meaningful integration into existing system landscapes.
• New AI models such as DeepSeek show that even intelligent models can be produced for relatively little energy and cost, and that the entry threshold for new developments is decreasing. This also presents opportunities to Switzerland in the competition for new AI.
• Highly complex, self-learning AI systems that generate new knowledge can continue to foster an energy-intensive digital arms race.

Zurich, 27 February 2025 – Large language models have fundamentally changed the way content is produced and consumed. But while companies and bloggers, for example, are increasingly opting for AI-generated content, there is growing fatigue with this flood of automated data. At the same time, artificial general intelligence (AGI) is drawing closer, which is capable of mastering every intellectual task at least as well as a human being. Nevertheless, companies are struggling to prepare their internal systems and data quality for this situation. By the time low-cost open-weight models such as DeepSeek appeared, there were signs that Switzerland, too, still has time to successfully position itself in the race for intelligent AI models. With the success of the small AI models, it is not unrealistic to assume that a cutting-edge model will soon be developed in Switzerland.

In its second position paper, the auditing and consulting firm EY Switzerland sheds light on the near future of artificial intelligence in three propositions.

Proposition 1: AI content fatigue on the rise

The digital landscape shows clear signs of content exhaustion. Users are becoming increasingly more adept at recognizing AI-generated content by its predictable patterns and overly polished tone, such as overly blunt formulations, generic statements and lack of depth. Content that is well-structured but lacks a real human perspective loses relevance. Many of these articles offer little more than repackaged information and rarely bring new aspects into the discussion.

Companies that use AI often program their AI agents to use smoothly polished corporate language to avoid potential controversies. Although some of these hurdles can be overcome with the right technology, today there is a flood of content that seems sterile and detached from human experience.

Adrian Ott, Partner and Chief Artificial Intelligence Officer at EY Switzerland, says: “Users are increasingly drawn to authentic human perspectives to understand the context and the person behind the opinions. The rise in AI-generated content has made it harder than ever to find original views.”

This development has led many users to look for alternative sources of information. Interviews, podcasts and traditional media formats are experiencing a renaissance, as they enable more authentic engagement with topics and tend to allow more transparency in respect of the origin of opinions. This shows that, at least for now, AI technologies can efficiently create content, but do not replace the human need for real insights and individual perspectives.

Use of AI with (human) judgment

When used wisely as an editorial and quality control tool, AI can enhance human creativity rather than replacing it. When creating a new article with AI, there are now a number of ways to completely remove unnecessary jargon and AI writing styles, keeping the content fresh and easy to read and allowing authors to focus on their original thoughts. An entire branch of science has emerged around the interaction between humans and AI: human-AI interaction (HAX). It investigates this link and suggests approaches for the future.

Proposition 2: Corporate data remains a challenge for artificial general intelligence (AGI)

Artificial general intelligence (AGI) theoretically surpasses humans in many cognitive abilities, but integrating it into companies remains a challenge. This is because corporate knowledge is based on data, processes and know-how accumulated over decades, which can often only be accessed by going directly to experienced employees. “The dream of loading all company data into AGI that then immediately gains an overview of the company will probably not work as expected,” says Ott. “The question is not only when AGI will come, but also how long it will take to integrate it meaningfully into our business world.”

Training AI systems on a complex network of company information is not only technically challenging, but also costly. Data protection concerns and regulatory hurdles further complicate the process. “Companies need to make sure that sensitive information does not get into neural networks, as these networks cannot forget what they have learned,” adds Ott.

Rather than a sudden revolution, it is to be expected that the integration of AI into operational processes will develop more gradually. The success of this transition will not be measured by how quickly companies can deploy AI systems, but by how effectively they can integrate them into their existing operations while maintaining reliability and trust.

Proposition 3: Efficient new AI models will reduce energy consumption

Another major issue in the field of AI productivity is the immense amount of energy that modern language models require. Large AI models, such as those developed by tech giants, require enormous computing capacity and thereby consume a large amount of power. As a result, more and more researchers and companies are looking for efficient AI models that can manage on fewer resources, yet still compete with the intelligence of large models.

New AI models such as DeepSeek, developed in China, show that powerful language models can also be trained with a fraction of the resources used previously. This raises the question of whether the need for giant, energy-intensive AI systems might be obsolete in the future.

In practice, the AI landscape is likely to develop along two lines: on the one hand, powerful, self-learning models that are intended to greatly expand the current state of knowledge of mankind and research new findings; and on the other hand, more efficient, more compact models for everyday use which make the newly generated knowledge available at low cost. Companies are faced with the decision of whether to invest in comprehensive, energy-intensive AI models in their specialist field or to opt for cost-effective models with a broad knowledge base.

“The race to improve AI capabilities is not just about applying existing knowledge correctly – it is about pushing the boundaries of what we currently know,” says Ott.

Conclusion: The human-AI partnership

There is growing fatigue with AI-generated content. In a time of information overload and AI bots, authenticity and transparency in respect of the origin of opinions and knowledge are once again in demand.

At the same time, even cautious companies that may have already experienced failures with AI solutions should prepare for exponential developments and act now to ensure that their data, processes and internal structures are up to speed with these rapid developments. And for highly specialized companies that have not yet considered developing their own AI models for cost reasons, the success of high-performance and cheaper models such as DeepSeek raises the question of whether specialized in-house development may actually give them a considerable competitive advantage and independence from third-party providers and other countries.

If you have any questions, please contact: daniele.mueller@ch.ey.com. 

Read the article: AI-generated content: challenges and opportunities

About the global EY organization

The global EY organization is a leader in assurance, tax, transaction and advisory services. We leverage our experience, knowledge and services to help build trust and confidence in the capital markets and in economies all over the world. We are ideally equipped for this task — with well trained employees, strong teams, excellent services and outstanding client relations. Our global purpose is to drive progress and make a difference by building a better working world — for our people, for our clients and for our communities.

The global EY organization refers to all member firms of Ernst & Young Global Limited (EYG). Each EYG member firm is a separate legal entity and has no liability for another such entity’s acts or omissions. Ernst & Young Global Limited, a UK company limited by guarantee, does not provide services to clients. Information about how EY collects and uses personal data and a description of the rights individuals have under data protection legislation are available via ey.com/privacy. For more information about our organization, please visit ey.com.

EY’s organization is represented in Switzerland by Ernst & Young Ltd, Basel, with 10 offices across Switzerland, and in Liechtenstein by Ernst & Young AG, Vaduz. In this publication, “EY” and “we” refer to Ernst & Young Ltd, Basel, a member firm of Ernst & Young Global Limited.