On June 29, the United States Federal Trade Commission (FTC) released a blog detailing the competition concerns surrounding Generative Artificial Intelligence. It mentioned that data, skilled talent, and computational resources are the building blocks of generative AI, and says that control over these building blocks can affect competition in the AI market. “As competition issues surrounding generative AI continue to develop, the Bureau of Competition, working closely with the Office of Technology, will use our full range of tools to identify and address unfair methods of competition,” it says.
Why it matters:
As generative AI becomes an integral part of the tech world, it is important to understand the kind of competition implications that could emerge from it. Anti-competitive practices can impose barriers to entry for smaller businesses that could otherwise create AI tools that would significantly improve an average person’s life.
How can AI development become anti-competitive?
- Access to data: To create new content, generative AI models have to be trained on vast amounts of data, and access to this data is much easier for bigger companies than for up-and-coming ones. According to the FTC, established companies (especially ones that run social media platforms) may benefit from already having access to vast quantities of data that they have collected from users over the years. Established companies are also more likely to have developed proprietary data collection tools which would give them an edge over new players. The competitive advantage of data access becomes even more pressing in industries such as healthcare and finance, where data is highly regulated and thus not easily accessible to new companies.
- Small talent pool: Generative AI development needs skilled engineers and researchers, and the FTC says that since the “requisite engineering talent is scarce, powerful companies may be incentivized to lock in workers and thereby stifle competition from actual or would-be rivals.” It says that individuals working in the AI space must be permitted to move freely and must not be hindered by non-compete agreements to ensure a competitive marketplace.
- Computational resources: The process of creating an AI model from scratch requires significant computational resources, which include specialized chips like graphical processing units (GPUs) that can be expensive to operate and maintain. While newer AI companies can meet this demand for computational resources through cloud computing services, even those can be expensive. Besides, cloud computing services also have other challenges. “There are reports, for example, that the spike in demand for server chips that can train AI has caused a shortage, prompting major cloud-server providers such as AWS, Microsoft, Google, and Oracle to “limit their availability for customers,” the FTC says.
- Control over adjacent markets:The FTC mentions that bigger companies that control adjacent markets could attempt to foreclose competition through bundling and tying. (quick context: bundling is the practice of selling multiple products as a package, and tying is the conditioning the sale of a product with the sale of another product) Bigger companies can link their generative AI with pre-existing offerings, thereby overshadowing the products that other, smaller companies are offering. Companies that offer both computational services and generative AI products could “use their power in the compute services sector to stifle competition in generative AI by giving discriminatory treatment to themselves and their partners over new entrants,” the FTC says. If a bigger AI company offers both – a generative AI model and API (application programming interface) for other businesses to leverage their AI, they may only offer it in ways that protect their dominant position.
- Consolidating the market: Bigger companies could also try to buy out nascent rivals instead of trying to out-compete them. They could also try to buy up critical applications necessary for the development of generative AI models to cut off their rivals’ access to them.
- First mover’s advantage: Companies that have been the first to create generative AI models and have had the chance to interact with multiple users would thus be able to create better results than rival products. This creates a positive feedback loop where users improve the functionality of a generative AI model, and this attracts more users to the model, which in turn, makes it harder for newer companies to break through into the market.
The pros and cons of Open-sourcing AI
The FTC highlights that open sourcing (making the original source code freely available so that the AI model can be modified and redistributed) opens up the playing field for just about anyone to develop, iterate on, and deploy AI models using smaller datasets and lower-cost consumer hardware. However, open source has its fair share of issues. The FTC says that sometimes, “Firms that initially use open-source to draw business, establish steady streams of data, and accrue scale advantages can later close off their ecosystem to lock-in customers and lock-out competition.”
Open-source code can also be misused by users who may bypass protections that were built into an AI model. The FTC mentioned that “while open-source AI image generation tools were released with built-in restrictions on the types of images that could be generated, malicious users removed these protections and utilized the models to create non-consensual intimate images.”
STAY ON TOP OF TECH NEWS: Our daily newsletter with top stories from MediaNama and around the world, delivered to your inbox before 9 AM. Click here to sign up today!
Also read:
- Five Talking Points: Open AI’s Plan To Launch An App Store For AI Models
- AI Companies Pushing For Regulation: Key Issues Discussed In The US Subcommittee Hearing On AI Oversight
- UK To Review AI Models For Competition, Consumer Protection Implications
