DOJ, FTC Divvy Up AI Antitrust Oversite
With the U.S. Congress seemingly too paralyzed to take meaningful action on artificial intelligence (or really anything else), federal regulators continue to try to step into the breach. Further to our previous post, the Department of Justice and the Federal Trade Commission are close to reaching an agreement on dividing up antitrust scrutiny of the largest AI companies, according to a report in the New York Times and subsequently confirmed by CNN.
Though not yet final, according to the reports, the agreement would see the FTC continue its current probe of OpenAI and Microsoft (along with Amazon and Anthropic) while DOJ will take on Nvidia and continue its oversight of Google. Separately (or perhaps not), the FTC expanded its investigation of Microsoft, seeking information on its recent acquihire and $650 million licensing deal with startup Inflection AI and whether the deal was deliberately structured to skirt federal antitrust review.
The agencies are concerned primarily with the highly concentrated structure of the AI market, where a handful of already dominant technology companies hold an effective monopoly on the financial and computing resources required to train and run large generative AI models. But at the risk of repeating myself, it’s clear from last week’s comments by Jonathan Kanter, head of DOJ’s antitrust division, and earlier comments by FTC chair Lina Kahn, both clearly have the effects of that concentrated structure on artists and rights owners on their radar.
The coordination between the agencies will be welcome news to those hoping for a robust regulatory response to the rapid and disorienting rise of generative AI technology. But there’s only so much law enforcement agencies can do using the laws currently on the books. Investigations and litigation, the agencies’ main tools, also take time. A lot of time. And it can lead to inconsistent or even conflicting results.
Microsoft, Google, Meta and Amazon, by now, are also quite familiar with regulatory scrutiny of their market power and M&A activity. They have all been there before and know the drill. And they know that the slow grinding of the legal process plays to their advantage.
Even aggressive action by DOJ and FTC against specific AI companies, moreover, will still leave the U.S. well short of other jurisdictions in developing a comprehensive regulatory and policy framework around AI, particularly the European Union.
For creators and rights owners, any interest by U.S. authorities to address their concerns with the unchecked advance of generative AI technology and the power of the companies driving it is welcome. Yet, while both DOJ and the FTC have begun to take up the cause, the theory of the case for using antitrust law to address what might otherwise fall into the domain of copyright and other IP laws is a novel one, not yet tested in court.
Publishers and other rights owners are nonetheless seizing on the agencies’ attention to their concerns while they have it. Last week, the president and CEO of the News|Media Alliance, Danielle Coffey, sent a letter to Kanter and Kahn to urge action against Google over its new search Overview feature that summarizes information drawn from publishers’ sites in response to search queries limiting users’ need to click through the original sources.
“In a competitive market, publishers can compete with Google for traffic and advertisers on the strength of their respective content. But Google has the requisite power to prevent this healthy competition by using its search monopoly to gain access to publishers’ content that allows it to keep traffic on Google and off publishers’ sites,” she wrote. “The only way for publishers to opt out of AI Overviews (i.e., Google’s GAI) and stop Google’s monopolistic practice is to effectively opt out of search distribution, which is not a viable option for publishers. Accordingly, no practical solutions exist for publishers seeking to protect their original content from Google’s misappropriation.”
Striking while the iron is hot.
This Week
Feeling the Heat
With all the heightened legislative jawboning and regulatory scrutiny of AI companies, Big Tech is clearly starting to feel the heat in Washington. This week, the Chamber of Progress, a tech industry coalition whose members include Amazon, Apple and Meta launched a new charm offensive called “Generate and Create” meant to highlight all the ways artists are using AI to ramp up their creative output and how AI lowers barriers to producing art by reducing the need to actually master an artistic skill. As with many Washington messaging campaigns, however, the subtext here is more important than the text. “The principle of fair use is well established over decades of copyright law,” said the group’s CEO, Adam Kovacevich, a former Google executive. “All art has responded to what’s come before…To the extent that human artists have always built art based on what came before, that’s what gen AI services are doing too.” Your mileage may vary.
WMG Gears Up for Growth
The folks at Warner Music Group must be feeling pretty flush these days. This week, it hired investment banker Michael Ryan-Southern away from Goldman Sachs to head up WMG’s M&A activity. “We have a clear strategy in expanding our offerings to serve more artists across a wider array of their careers. And we are building against that.” WMG CEO Robert Kyncl said. “We have our team working really hard, building all the right features that we need. We always look at ways to accelerate because all of this work takes time.” Nice to see the music business working well for somebody.
Eroding Interest in Content ID?
YouTube’s Content ID system processed over 1 billion copyright claims in the final six months of 2023, according to the company’s latest report, the highest volume the automated content recognition system has fielded over a comparable period of time to date. But the number of rights owners actually using the system has been shrinking. Access to Content ID has always been limited to a small number of large IP owners. But the number using the system in the last six months last year fell to 7,791, compared with 8,900 in the first six months, a drop off of more than 12%. Among those, only 4,511were actively using the system last year. No reason was given for the decline, but theories are welcome.