Reporting notes from the AI trenches: Jan 1-7
The NYT lawsuit against Microsoft/OpenAI for copyright infringement; a voice cloning AI startup opening up to political ads; and what else I'm digging into
Happy new year 2024!
All was quiet on the AI front when I got back to work reporting for VentureBeat the day after New Year’s last week. So much so, in fact, that I wondered why my email inbox was suddenly so empty.
Of course, this was only a brief pause — many AI industry folks are headed to Vegas for CES this week; the World Economic Forum’s annual conference in Davos is coming up; and most of us were just wading through an inbox that was filled with news from before, and during, the holiday season.
The biggest AI news between Christmas and New Year’s was inarguably the lawsuit filed by the New York Times against OpenAI and Microsoft for copyright infringement on December 27.
I tackled this in an AI Beat column last week — in which I maintained that no matter which side you think will win the battle, there is no doubt that the NYT lawsuit will be one of the most important copyright fights to watch in 2024.
The lawsuit maintains the defendants should be held responsible for “billions of dollars in statutory and actual damages” related to the “unlawful copying and use of The Times’s uniquely valuable works.” It also calls for OpenAI and Microsoft to destroy any chatbot models and training data that use New York Times copyrighted material.
The case is so important for 3 key reasons:
Legal experts have long predicted that a case around AI and copyright could make it to the Supreme Court — and a high profile case like the New York Times vs. Microsoft and OpenAI could be just the ticket to a date with the nine justices.
Industries affected by generative AI are not taking a wait-and-see approach. For example, the New York Times remembers the famous Authors Guild v. Google, a federal copyright case that centered on whether Google Books’ effort to digitize books constituted “fair use,” and wound its way through the courts for a decade before being decided in Google’s favor in 2015.
Both sides of the copyright issues in this case have strong arguments and big-time legal representation.
Check out what I had to say on the NYT vs. Microsoft/OpenAI case on this episode of VICE Motherboard’s Cyber podcast.
Last week I also dug deeper into the use of AI around election misinformation — this time a startup that has expanded its voice cloning offering to political ads.
You might remember that before the holidays, I wrote a VentureBeat story, “AI will make 2024 US elections a ‘hot mess,’” tied to my interview with machine learning researcher
of .As I was hunting for a good post-New Year’s story, I got a reach-out from a startup, Instreamatic, an AI ad tech company with voice cloning audio/video ad capabilities that was branching out to what the PR rep called the ‘wild, wild (did I mention wild?) world of political ads.
The company shared a demo of a video campaign that adapts to changing events without a candidate recording new audio:
I spoke to Instreamatic’s CEO and co-founder Stas Tushinskiy for the story, who insisted the company has guardrails built in to make sure its product is not used for election disinformation.
As someone who has covered adtech and marketing tech for over a decade, I can definitely see how the Instreamatic product could help with what has long been a tedious, manual ad-producing process. But in the context of fears around the use of AI in the 2024 elections — for example, there are currently no federal rules for the use of AI-generated content, such as ads, in political campaigns — I wonder what it means for the future of political advertising, even if it just adds to public confusion and mistrust.
Here’s what else I spent time digging into last week:
Follow-up reporting on this VentureBeat story: The widening web of effective altruism in AI security — keep your eyes peeled for a new article from me later this week.
Looking into enterprise distribution of AI tools like Microsoft Copilot — h/t to
on this one — as well as how enterprise customers play into the latest OpenAI and Anthropic valuation claims.
As always, if you have any tips or hints or comments or questions for me, please feel free to email me (sharon dot goldman at venturebeat dot com) or comment/message here on Substack.
👋 Sharon