The over-hype of generative AI vs actual use: a study

Public awareness and use of generative AI

Generative AI, particularly ChatGPT, is widely recognised, with around 50% of the online population across the six surveyed countries having heard of it. However, frequent use is rare. Daily usage rates are as low as 1% in Japan, 2% in France and the UK, and 7% in the USA. Most individuals who have used generative AI have done so only once or twice, indicating that it has not yet become a routine part of internet use.

Detailed findings

• Awareness: Despite widespread awareness, 20-30% of the online population in the surveyed countries have not heard of popular AI tools.
• Usage: ChatGPT is the most widely used generative AI tool, significantly more than Google Gemini and Microsoft Copilot.
• Demographics: Younger individuals (18-24 years) are more likely to use generative AI regularly, with 56% having used ChatGPT at least once, compared to 16% of those aged 55 and over.
• Purpose: Generative AI is used equally for obtaining information (24%) and creating various media (28%). Only 5% use it for news updates.

Public opinion on AI's impact across sectors

The survey indicates that most people expect generative AI to significantly impact various sectors within the next five years. However, trust in the responsible use of AI varies widely:

• Trust levels: Half of the respondents trust scientists and healthcare professionals to use AI responsibly, while less than one-third trust social media companies, politicians, and news media.
• Sector impact: Younger respondents are more likely to expect a significant personal impact from generative AI than older individuals.
• Quality of life: A plurality in four of the six countries believe AI will improve their lives, though many remain uncertain or pessimistic.

Sector-specific optimism and pessimism

There is considerable optimism for AI’s impact on science, healthcare, and entertainment. However, not all sectors are seeing such levels of optimism. Concerns prevail regarding AI’s effect on the cost of living, job security, and news quality, with more people pessimistic than optimistic, some seeing it as a catalyst or threat for these problems.

AI in journalism

Generative AI’s role in journalism is particularly contentious. The public expects AI-generated news to be less trustworthy and transparent but acknowledges that it could be more up-to-date and cheaper to produce.

Key findings:

• Journalists' use of AI: A significant portion of the public believes journalists use AI for tasks like editing spelling and grammar (43%), writing headlines (29%), and drafting articles (27%).
• Human oversight: About one-third of respondents think human editors check AI outputs before publication.
• Comfort levels: People are generally more comfortable with human-generated news. However, there is slightly more acceptance of AI-generated content for soft news (e.g., fashion and sport) than for hard news (e.g., international affairs and politics).
• Labelling: The majority want AI-generated news to be clearly labelled, especially for writing the text of articles and data analysis.

Conclusion

The study highlights a significant gap between the hype surrounding generative AI and its actual integration into everyday life and various sectors. While there is broad awareness and sporadic use, frequent use remains limited, and public opinion is mixed regarding AI’s future impact. In journalism, particularly, trust and comfort with AI-generated content are low, underscoring the need for transparency and human oversight. As AI technology continues to evolve, addressing these concerns will be crucial for its broader acceptance and responsible implementation.