SHENZHEN, China, Aug. 7, 2024
/PRNewswire/ -- AI video creation is a key development focus in the
generative AI industry. Recently, Kling Video AI Model,
published by Kwai, took the lead in launching large-scale open user
testing, triggering user trials and a surge of "creative-oriented"
AI video content on content platforms. Users' response has been
positive, injecting a sense of novelty among platform users. The
Lumière brothers' silent black-and-white film, The Arrival of a
Train, premiered a century ago, creating a stunning effect on
audiences. Today, a century later, with the support of AI video
models, AI will pave a path for "realizing imagination" in content
development. A new wave of "AI Short Videos" development may
emerge, highlighting the convergence of conventional and
unconventional, virtual and real, and traditional and modern
approaches. Given the diminishing marginal utility of traditional
content transformation, this could become a new direction for
content creation and provide significant opportunities for AI to
break through.
I. AI Video Models' Delayed Arrival, Kwai Kling Takes
the Lead To Reach a New Level
As a track in generative AI, AI video creation has garnered
attention and significant investment from various sectors.
According to incomplete statistics, over 30 AI video models and
tools exist in China alone.
However, despite the vast potential demand for AI text-to-video,
many products are still in the early stages of technical
exploration, with the usability of deployed models generally
needing improvement. At the beginning of the year, Sora unveiled
AI-generated videos that were up to one minute long, igniting
market and user enthusiasm. However, the delayed opening of user
testing led to a downturn in market sentiment towards AI
text-to-video. This includes evaluations from some foreign
producers and media professionals during Sora trials, causing the
market to adopt a wait-and-see attitude towards text-to-video
technology.
Additionally, some available AI video models face high trial
costs, reducing user interest in continuous experimentation. For
instance, Runaway's latest Gen3 version, despite its
internationally leading generation quality, comes at a steep price
of $12/month for just 125 seconds of
qualified output per month. Against this backdrop, achieving
low-cost usage thresholds, minute-level output, and basically
usable generation quality remains a significant market expectation
for the generative AI industry.
Amidst the notable development pain points in AI video
generation, Kwai Kling emerged, taking the lead in large-scale user
trials. Launched at the beginning of last month, the Kwai Kling
provides both web and mobile access points, with the mobile version
requiring users to complete tasks within the Kwaiying App. It
opened application-based testing, with numerous users already
granted trial access and actively engaging in content creation. The
model supports various AI video creation methods, including
text-to-video, image-to-video, and video continuation, focusing
more on user experience. According to data on SimilarWeb, after
opening on June 6, the overall
website UV for the Kling showed an upward trend. Entering July, the
web UV continued to increase, with peak values approaching 100,000
DAU. On the mobile side, since both application and
generation model are completed within the Kwaiying App, MoonFox
iApp data reveals that from the second week of June, the weekly
average DAU of Kwaiying App increased significantly.
Particularly after the first week of July, the weekly average
DAU increased by about 1 million compared to early
June, indicating strong market enthusiasm for Kling
Changes in Monthly
Active Users of Kwai Kling - Web Version in June-July
(Partial)
|
Date
|
Weekly Average
DAU
|
Note
|
6.06
|
41043
|
Kwai Kling AI Video
Model opens for testing
|
6.13
|
43028
|
|
6.25
|
56508
|
Kwai Kling AI Video
Model starts trending on social platforms
|
7.06
|
84554
|
|
Data Source: SimilarWeb; Data period: June 2024
Changes in Weekly
Average DAU of Kwaiying App in June - July
|
Date
|
Weekly Average
DAU
|
Note
|
10-16 June
|
5.401
(million)
|
Kwai Kling AI Video
Model opens for testing
|
17-23 June
|
5.421
(million)
|
|
24-30 June
|
5.884
(million)
|
|
1-6 July
|
6.450
(million)
|
|
Data Source: MoonFox iApp; Data Period: June 2024 - July
2024
II. Empowered by AI Video Models, Young Generations Spark a
New Wave of "AI Creative-Generated" Content, Injecting a Novel
Experience for Users
Following the launch of the Kling AI Video Model, "AI
Creative-Generated" videos have recently gained popularity on video
platforms. For instance, various themed AI video content has
appeared on the Bilibili platform, with some videos surpassing
millions of views and thousands of likes. These contents are mostly
highly imaginative creations. The users contributing to these
creative AI productions will likely be predominantly from younger
user groups. According to data on SimilarWeb, over 50% of web users
of Kling AI Video Model are under 34 years old, with the 18-24 age
group accounting for nearly 20%.
User Composition of
Kwai Kling AI Video Model - Web Version
|
Gender
Distribution
|
Percentage
|
Male
|
72.7 %
|
Female
|
27.3 %
|
Age
Distribution
|
|
Aged 18-24
|
19.0 %
|
Aged 25-34
|
36.0 %
|
Aged 35-44
|
21.0 %
|
45 years old and
above
|
24.0 %
|
Data Source: SimilarWeb; Data Period: June 2024
From the user feedback and evaluations perspective, enthusiasm
for the Kling AI Video Model is currently high. We analyzed social
media user comment data related to the Kling AI Video Model,
excluding official accounts, influencers, and KOLs, retaining only
ordinary user comment data. We found that overall, users hold a
positive attitude towards the current Kling AI Video Model. User
emotions are mainly concentrated in four categories: curiosity,
excitement, anticipation, and satisfaction. Users exhibit high
levels of pleasure and arousal towards the Kling AI Video Model.
The high arousal level also indicates that compared to current
content types, the Kling AI Video Model is more capable of
activating users' psychological enthusiasm and willingness to act,
suggesting that more users will be willing to try using and viewing
related content in the future.
Emotional Analysis of
Kwai Kling AI Video Model User Evaluations
|
Emotion Type
|
Percentage
|
Curiosity
|
45.0 %
|
Excitement
|
31.8 %
|
Anticipation
|
12.7 %
|
Satisfaction
|
5.5 %
|
Wait-and-see
|
9.1 %
|
Data Source: MoonFox Big Data Analysis; Data Period:
June 2024
III. The Concept of "AI-Generated Short Videos" May Break
Through in the Future, Presenting a Different Development Approach
from Traditional Content Creation
How AI generation can achieve true implementation has been a key
concern across various sectors. Due to factors such as
hallucination issues, many current AI applications still face
retention problems. As a subdivision within this field, we believe
AI video creation has the potential to gain further user
recognition on the usage end in the future.
Drawing from the experience of the previous round of PC and
internet development, the focus on AI is not just about the
efficiency revolution but also about realizing imagination. Taking
QQ as an example, the internet-era QQ provided a vast space for
imagination, such as QQ Space, QQ Show, emoticons, and other
creative designs. Compared to traditional content at the time, the
experience was quite exceptional. Behind online social interaction,
there was actually a large number of novel content experiences
supporting it, not just simple efficiency improvements. Similarly,
current AI virtual social interaction applications with daily user
engagement times of several dozen minutes follow the same
principle. On the other hand, entertainment content, compared to
efficiency tools, inherently has broader coverage and stronger user
stickiness in terms of experience.
However, in the AI era, we believe that the "novelty" to be
embodied in video content creation differs from traditional
interpretations. In a situation where user interests are relatively
stable and user interest circles are very distinct, breakthroughs
are necessary to refresh the experience. Specifically, we believe
there might be three aspects: The fusion of conventional and
unconventional, the integration of virtual and real, and the blend
of traditional and modern, thereby integrating with the real world
to capture user interest. Secondly, while the current high cost of
AI video creation is a fact, using AI for large-scale generation of
long videos may not be the short-term implementation direction.
However, generating short videos or integrating with existing video
segments is quite possible. AI short videos might become a
direction where AI video models can be practically implemented in
the near future.
A century ago, the earliest silent black-and-white films were
very limited in duration due to cost constraints at the time. For
example, the Lumière brothers' The Arrival of a Train was
only 50 seconds long. However, when it premiered, its realistic
portrayal left a deep impression on the entire audience, thus
propelling a massive film industry. A century later, the AI era may
potentially guide creators to explore more ideas that are difficult
to realize with traditional content, opening up a new approach to
video creation, which could have profound implications for future
user content consumption.
MoonFox Data is a leading expert in all-scenario data insights
and analytics services. With a comprehensive, stable, secure and
compliant mobile big data foundation, as well as professional and
precise data analysis technology and AI algorithms, MoonFox Data
has launched iAPP, iBrand, iMarketing, Alternative Data and
professional research and consulting services of MoonFox Research,
aiming to help companies gain insights into market growth and make
accurate business decisions.
About us: Aurora Mobile (NASDAQ:
JG) established in 2011, is a leading customer engagement and
marketing technology service provider in China. Its business includes notification
services, marketing growth, development tools, and data products.
As its sub-brand, MoonFox Data is a leading expert in data insights
and analysis services across all scenarios, aiming to help
companies gain market insights and empower precise
decision-making.
- Our Information:
Website: https://www.moonfox.cn/en
Contact number: 400-888-0936
- Contact us:
Name: Felix
Title: Director of Sales, Industry Insight Division
Tel: +86 -13366276383
Email: zhouzt@jiguang.cn
Address: 608/F, Tower B, Wintrust Center, No.
1 Xidawang Road, Chaoyang District,
Beijing, China
Name: Janette Zhou
Title: Marketing Manager
Tel: +86-13928239332
Email: zhouxt@jiguang.cn
Address: 608/F, Tower B, Wintrust Center, No.
1 Xidawang Road, Chaoyang District,
Beijing, China
View original
content:https://www.prnewswire.com/news-releases/moonfox-analysis--the-late-entry-of-kwais-kling-can-ai-powered-short-video-become-its-new-breakout-opportunity-302221741.html
SOURCE Aurora Mobile Ltd