The Achilles heel of AI startups: no shortage of money, but a lack of training data
With ChatGPT driving the AIGC boom rapidly around the world, a large number of
generative AI startups are springing up right now.
However, even though
these startups can easily obtain billions of dollars of investment funds, there is
still a fatal weakness that is almost unavoidable at the moment - the lack of
training data, which may eventually become the biggest "stumbling block" on the road
to success for these startups.
Brad Svrluga, co-founder and general partner
of venture capital firm Primary Venture Partners, said they have received a number
of pitches from AI startups. These AI startups are likely pursuing the creation of
great AI applications, but many of them don't have access to the training data that
would allow them to build robust applications, not to mention the proprietary data
that would help them build a competitive moat in their business.
Data is harder to come by than money
According to PitchBook,
the size of venture capital investment in generative AI startups has grown from $4.8
billion in 2022 to $12.7 billion in the first five months of 2023.
Many of
these companies are now looking to build more niche AI models in areas such as
finance or healthcare, but getting training data sets for those areas isn't easy.
Paul Tyma, chief technology officer at Bullpen Capital, then pointed out
that building actual models is now somewhat commoditized, and that the real value
lies in the data.
Some AI startups are aiming to work with large, data-rich
companies. But large companies also worry about their data being used to train
external models and wonder what will happen - after all, the data is part of their
intellectual property. Of course, startups can address intellectual property issues
by simply using their customers' data and training different models for each
customer.
Andy Wilson, co-founder and CEO of legal tech firm Logikcull,
noted that it's also a challenge to convince companies that you have strong
cybersecurity capabilities and can actually protect that data.
Large
enterprises hold a huge advantage
Large technology companies
clearly have an advantage over startups when it comes to artificial intelligence
applications and data processing, in part because they have gained the trust of
large customers who feel more comfortable with their handling of data.
Founded
in May 2003, Knowlesys is a high-tech company that specializes in Internet
information processing. It has established rich experience and competitive
advantages in the fields of accurate monitoring and vertical mining of mass
information on the Internet.
Knowlesys Intelligence System is developed
based the world's leading open source intelligence extraction technology, with the
advantages of quick identification and full coverage. It enables users to monitor
the whole Internet, including social media(Twitter, Facebook, YouTube,
Instagram...), traditional websites (Forum, Chat Room, News...) and dark web in a
timely manner.