On Dec. 9, OpenAI made its artificial intelligence video generation model Sora publicly available in the U.S. and other countries.
Cfoto | Future Publishing | Getty Images
The U.K. is drawing up measures to regulate the use of copyrighted content by tech companies to train their artificial intelligence models.
The British government on Tuesday kicked off a consultation which aims to increase clarity for both the creative industries and AI developers when it comes to both how intellectual property is obtained and then used by AI firms for training purposes.
Some artists and publishers are unhappy with the way their content is being scraped freely by companies like OpenAI and Google to train their large language models — AI models trained on huge quantities of data to generate humanlike responses.
Large language models are the foundational technology behind today’s generative AI systems, including the likes of OpenAI’s ChatGPT, Google’s Gemini and Anthropic’s Claude.
Last year, The New York Times brought a lawsuit against Microsoft and OpenAI accusing the companies of infringing its copyright and abusing intellectual property to train large language models.
In response, OpenAI disputed the NYT’s allegations, stating that the use of open web data for training AI models should be considered “fair use” and that it provides an “opt-out” for rights holders “because it’s the right thing to do.”
Separately, image distribution platform Getty Images sued another generative AI firm, Stability AI, in the U.K., accusing it of scraping millions of images from its websites without consent to train its Stable Diffusion AI model. Stability AI has disputed the suit, noting that the training and development of its model took place outside the U.K.
Proposals to be considered
First, the consultation will consider making an exception to copyright law for AI training when used in the context of commercial purposes but while still allowing rights holders to reserve their rights so they can control the use of their content.
Second, the consultation will put forward proposed measures to help creators license and be remunerated for the use of their content by AI model makers, as well as give AI developers clarity over what material can be used for training their models.
The government said more work needs to be done by both the creative industries and technology firms to ensure any standards and requirements for rights reservation and transparency are effective, accessible and widely adopted.
The government is also considering proposals that would require AI model makers to be more transparent about their model training datasets and how they’re obtained so that rights holders can understand when and how their content has been used to train AI.
That could prove controversial — technology firms aren’t especially forthcoming when it comes to the data that fuels their coveted algorithms or how they train them up, given the commercial sensitivities involved in revealing those secrets to potential competitors.
Previously, under former Prime Minister Rishi Sunak, the government attempted to agree a voluntary AI copyright code of practice.
AI copyright rules: U.K. versus U.S.
In a recent interview with CNBC, the boss of app development software firm Appian said he thinks the U.K. is well placed to be the “global leader on this issue.”
“The U.K. has put a stake in the ground declaring its prioritization of personal intellectual property rights,” Matt Calkins, Appian’s CEO, told CNBC. He cited 2018’s Data Protection Act as an example of how the U.K. is “closely associated with intellectual property rights.”
The U.K. is also not “subject to the same overwhelming lobbying blitz from domestic AI leaders that the U.S. is,” Calkins added — meaning it might not be as prone to bowing down to pressure from tech giants as politicians stateside.
“In the U.S., anybody who writes a law about AI is going to hear from Amazon, Oracle, Microsoft or Google before that bill even reaches the floor,” Calkins said.
“That’s a powerful force stopping anyone from writing sensible legislation or protecting the rights of individuals whose intellectual property is being taken wholesale by these major AI players.”
The issue of potential copyright infringement by AI firms is becoming more notable as tech firms are moving toward a more “multimodal” form of AI — that is, AI systems that can understand and generate content in the form of images and video as well as text.
Last week, OpenAI made its AI video generation model Sora publicly available in the U.S. and “most countries internationally.” The tool allows a user to type out a desired scene and produce a high-definition video clip.
A Boeing 767-332(ER) from Delta Air Lines takes off from Barcelona El Prat Airport in Barcelona, Spain, on October 8, 2024.
Joan Valls | Nurphoto | Getty Images
CrowdStrike moved Monday evening to dismiss Delta Air Lines’ lawsuit around the July cybersecurity outage that led to canceled flights and stranded passengers, arguing that the airline’s litigation was an attempt to circumvent the contract between the two companies.
The agreement between CrowdStrike and Delta includes a clause limiting CrowdStrike’s liability and a cap on damages, which the cybersecurity provider says Delta is now trying to skirt. CrowdStrike also argued in its filing that Georgia law prevents Delta from converting a breach of contract into tort claims.
“As an initial matter, Georgia’s economic loss rule specifically precludes Delta’s efforts to recover through tort claims the economic damages it claims to have suffered,” CrowdStrike wrote.
Delta said the July cybersecurity outage cost the company more than $500 million in canceled flights, refunds and passenger accommodations. It is seeking to recoup those costs from CrowdStrike through the suit. But the damage done to Delta’s reputation as a premium carrier can’t yet be quantified, nor has the impact of a Department of Transportation investigation into Delta over the outage.
Delta continues to rely on CrowdStrike services following the outage, likely because it is extremely difficult to change cybersecurity providers in systems as large and complicated as Delta’s.
Still, CrowdStrike said it moved quickly to try and help Delta — offers the cybersecurity company says were rebuffed. “We are good for now,” one message from a Delta executive cited by CrowdStrike read. The cybersecurity company said its executives were in close contact on the day of the outage.
“Delta repeatedly rebuffed any assistance from CrowdStrike or its partners,” CrowdStrike wrote.
CrowdStrike further argues that Delta’s own practices and systems led to the widespread delays and cancellations, unlike other industry peers who recovered much more quickly from the outage.
“Delta was an outlier. Although Delta acknowledges that it took just hours—not days—for Delta employees to” remediate the outage, CrowdStrike wrote in its filing, “cancellations far exceeded the flight disruptions its peer airlines experienced.”
The cybersecurity company’s stock took a sharp hit after the outage, plunging 44%. It’s since largely recovered from those losses, posting strong quarterly results even after lowering its guidance due to the incident. CrowdStrike has been helped by the relative stickiness of its products, especially at large enterprises.
A Delta spokesperson was not immediately available for comment.
Dexcom on Tuesday announced an artificial intelligence feature for its Stelo continuous glucose monitor that gives users a personalized look into how meals, sleep and activity impact their glucose levels. It’s the first iteration of a new generative AI platform that the company has been building with Google Cloud.
Stelo is an over-the-counter CGM that pokes through the skin to measure real-time blood sugar levels. The sensor launched in August and can be used by any adult who doesn’t take insulin.
The report reflects Dexcom’s effort to make Stelo more personalized and engaging for consumers as it works to penetrate a new market.
“The No. 1 feedback we get is users want to see more,” Jake Leach, chief operating officer at Dexcom, told CNBC in an interview. “They’re making an investment and wearing the product, and they want to be able to take the most advantage of all the data that they’re generating.”
Dexcom is using Google’s Gemini models and its Vertex AI platform as the foundation for its new AI offering. Vertex AI allows developers to build applications that synthesize different types of data, which can be notoriously challenging in health care.
Leach said Dexcom is also exploring how its generative AI platform can be used across its other CGM products, but the company is proceeding extra carefully since patients rely on them to prevent medical emergencies.
“It really felt like Stelo was the right place to do this for the first time,” he said.
An existing insights report has already been available to users within the Stelo app, but it followed a more standard template format each week. Dexcom believes the AI-generated report will be more valuable to users since it’s personalized, Leach said.
If there’s a week where a user is not moving enough after meals, for instance, the report would include relevant tips and educational materials to help.
Stelo’s AI reports don’t give users medical advice, though Dexcom has been using an AI framework from the U.S. Food and Drug Administration to help guide the feature’s development, Leach said. The FDA approved Stelo in March.
Eventually, Dexcom wants to use its generative AI platform to deliver real-time feedback to users instead of just weekly reports. The company is also exploring how the technology could act as a predictive indicator for potential problems, much like a check engine light on a car.
“It gives you a sense for what could be going on, and recommendations of where you might want to go to seek more advice,” Chris Sakalosky, vice president of strategic industries for Google Cloud, told CNBC in an interview.
Dexcom’s updated weekly report began rolling out to Stelo users this week.
Close-up of Databricks company logo on building facade, Rincon Hill, San Francisco, June 7, 2024.
Smith Collection/ gado | Archive Photos | Getty Images
Databricks, one of the most valuable privately held companies, announced a $10 billion financing on Tuesday that values the software maker at $62 billion.
With the money, Databricks will be able to provide liquidity to current and former employees, make acquisitions, and expand overseas, according to a statement. The company’s new valuation is up from $43 billion in 2023. Rival Snowflake was worth about $57 billion at Monday’s close.
Databricks sells software for analyzing and cleaning up data, and it also runs artificial intelligence models for clients. The software is available on the Amazon, Google and Microsoft clouds, which are also competitors.
The company expects to generate positive free cash flow for the first time with a $3 billion revenue run rate in the quarter that ends on Jan. 31, Databricks said. The company’s revenue in the October quarter grew more than 60% year over year.
Investors in the financing, of which it has raised $8.6 billion to date, include Thrive Capital, Andreessen Horowitz, DST Global, GIC, Iconiq Growth, Insight Partners, MGX, Sands Capital, WCM Investment Management and Wellington Management.
Technology investors have been anticipating a Databricks initial public offering for years. They may only have to wait a few more months.
ServiceTitan, a company with software for plumbers and others working in the trades, raised about $625 million in an initial public offering last week, and some investors have predicted that tech IPOs will become more frequent again in 2025 after a relative drought since late 2021.
Databricks did not offer new information about its expectations for an IPO on Tuesday.
“If we were going to go, the earliest would be, let’s say mid next year or something like that,” Ali Ghodsi, Databricks’ co-founder and CEO, said at the Cerebral Valley AI Summit in November.
Late-stage investors with large funds don’t have many options for what to back, Ghodsi said.
“There’s nowhere to put it, really, except maybe Databricks, Stripe or, you know, maybe OpenAI,” Ghodsi said.
Databricks made its fourth appearance on CNBC’s Disruptor 50 list of private companies in 2024.