Taking on big tech.
The issue is becoming so popular it's bringing together political adversaries like Donald Trump and Nancy Pelosi. Even Elizabeth Warren and Ted Cruz. Last week, the House Judiciary Committee announced it would be launching a bipartisan antitrust investigation into companies like Google, Facebook and Amazon.
Each of those tech giants has become enormously powerful, particularly as it relates to gathering personal data and influencing behavior. Increasingly, that control is being driven by AI & machine learning technologies—e.g., Google's Assistant and YouTube algorithms; Facebook's content flagging, filtering and moderation; and Amazon's Alexa, purchasing recommendations and AWS tools.
It's clear that AI is no longer a nascent prototype tech of the future. It's being industrialized and commercialized at a massive scale, impacting billions of people at the behest of the world's biggest companies.
"Essentially as AI/ML technology becomes more readily available, these huge firms are positioned to dominate and potentially be extremely hard to compete with—especially within certain core competencies," said PitchBook Emerging Tech analyst Cameron Stanfill, who recently covered the space in depth in his research report. "It's a look at what's to come and how central AI/ML is going to be for essentially all internet users and enterprises alike."
Stanfill added that the tech giants have all made it clear that implementing AI/ML throughout their business and product offerings and by running open source frameworks—like TensorFlow and PyTorch—that help build an ecosystem of development around their platforms, creating a moat of sorts and attracting AI/ML talent.
Venture capital investors, however, are still making plenty of bets on smaller players trying to compete in the space, perhaps by carving out tangential or niche areas where the giants aren't as firmly developed. According to PitchBook data, deal flow into US-based AI/ML startups has increased unabated for about a decade.
Here are a few ways the big tech companies are winning the AI race:
The sheer quantity of data mega-firms are dealing with allows these hyperscalers to have a clear edge, as AI/ML applications depend on massive datasets to help train models.
The prevalence of high-performance computing and specialized chips utilized by the huge firms—Google's TPUs, for instance—allows these companies to tackle problems and conduct research that would be impossible for competitors with lesser resources.
Another example is neural architecture search—i.e., algorithmically searching for new designs of neural networks instead of relying on human data scientists—which tends to require a lot of brute force computation. That gives the big tech companies an advantage in improving their models, given they can use expensive, state-of-the-art computers to iterate.
Indeed, the giants are getting bigger. And with so many AI/ML advantages in place to continually improve their large-scale networks, it's not hard to see why regulation looms.
Featured image via archy13/iStock/Getty Images Plus
Download an executive summary of our AI & machine learning report from PitchBook Emerging Tech Research.
If you have any questions or feedback on the research, we encourage you to email email@example.com.
PitchBook clients can get more information on the premium research by reaching out to PBPaidReports@pitchbook.com.