WTF now?!: Monday 27 January–Sunday 2 February 2025
Elliott Bledsoe
I am the grumpy cat of arts marketing! ?? ATM doing stuff w/ BlakDance, CircuitWest + others. President of Wikimedia Australia and Co-lead of Creative Commons Australia. On Turrbal/Jagera land
My weekly reading round-up
Chinese AI has Silicon Valley spooked while copyright and AI issues continue, internet filtering to reduce copyright piracy re-emerges in the US and Spotify talks up its importance in music while reducing music royalties.
As often happens, AI dominated the news cycle again this week. When it comes to DeepSeek, you may be asking yourself, WTF now?! And you aren’t alone. The Chinese AI player rivaling the best of Silicon Valley seemingly came out of nowhere. I look at what it is and why it has the US AI industry worried.
Also, the US Copyright Office reaffirms the importance of human creativity if AI outputs are to be protected, blocking foreign piracy websites is back on the agenda in the US and bundled Spotify subscriptions are impacting music royalties, even as the streaming giant boasts about its importance to the music industry.
What’s going?on?
Here’s what I’ve noticed this week:
DeepSeek has the US AI industry scared
DeepSeek has low cost, efficiency and being open source going for it, but being Chinese made is a big hurdle.
The AI space is fast moving and is a disruptive force, but the formal release of the DeepSeek-R1 model this week shows that even AI can be disrupted. You can be forgiven for having never heard of DeepSeek until recently. Their first major reasoning model – DeepSeek-V3 – was released on Christmas Day. That started a bit of a buzz. Since then, there has been plenty of praise and raised eyebrows in the industry for the Chinese AI maker. Its AI tools are absolutely viable contesters for any Silicon Valley AI system currently on the market.
The release of R1 during the week upped the ante, sending the US AI and finance sectors into a spin. The previously obscure Chinese player shot to the top of the Apple and Android app stores’ free app charts, quickly raising awareness and the userbase. DeepSeek has reportedly reduced the cost to develop AI systems significantly. This is a good thing for AI: efficiency lowers the cost of AI for new entrants to the market and for users, innovation pushes the field forward, less GPU demand can reduce the environmental impact of AI and it gives hope that AI developed outside the US can survive. But naturally, all that has the US AI industry, and its financiers, spooked.?
As Kylie Robison and Elizabeth Lopatto said on The Verge:
"It took about a month for the finance world to start freaking out about DeepSeek, but when it did, it took more than half a trillion dollars — or one entire Stargate — off Nvidia’s market cap. It wasn’t just Nvidia, either: Tesla, Google, Amazon, and Microsoft tanked."
The US has been quick to respond. While some commentators are expressing how impressive the newcomer is, others are more threatened. A growing list of US government agencies have banned or discouraged use of DeepSeek, reportedly including Congress, the US Navy, The Pentagon and NASA. The Trump administration has called it a “wake-up call” for the US industry and have instructed the National Security Council (NSC) to investigate the potential national security implications – the same reasoning behind the TikTok ban.
Also, OpenAI and Microsoft have been looking at whether DeepSeek used OpenAI’s API to exfiltrate data using developer accounts. Open AI claims evidence of distillation – a common technique that sees a smaller model trained on extracted data from larger, more capable models. If it is true, those accounts will be in breach of OpenAI’s terms of use. That didn't stop Microsoft from adding it to the Azure AI Foundry though.
The irony of OpenAI calling out DeepSeek for using its data to build an AI model is huge. The company built it’s AI dominance by training its models on the entire written content on the internet. And it didn’t tell anyone it was doing it or ask anyone’s permission to do so. We are still waiting to find out definitively whether OpenAI’s actions were legal under the US fair use copyright exception. Even if it was, the ethical quandry is bigger than the irony itself.
Speaking of ethics, they are complicated for DeepSeek too: On the one hand its models are partially open source, while on the other it censors some types of responses because of its Chinese origin. There are a range of things its models won't respond to, many of which are tied to political ideology, the image of the Chinese Communist Party and sensitive topics such as Taiwan’s democracy and the Tiananmen Square massacre. Plus user data is stored on servers in China, raising privacy concerns.
What’s worth reading on DeepSeek’s sudden rise:
Kylie Robison and Elizabeth Lopatto – Tuesday 28 January 2025
The Verge
Jess Weatherbed – Wednesday 29 January 2025
The Verge
Emmet Lyons – Wednesday 29 January 2025
CBS News
Ben Shepherd – Friday 30 January 2025
Signal
The US Copyright Office clarifies when AI outputs are protected by copyright
US Copyright Office report emphasises human expression for AI outputs to be protected.
The US Copyright Office has released the second part of its report on the legal and policy issues related to copyright and AI. This part looks at whether generative AI outputs are able to be protected by copyright under US law. It concludes that:
“… the outputs of generative AI can be protected by copyright only where a human author has determined sufficient expressive elements. This can include situations where a human-authored work is perceptible in an AI output, or a human makes creative arrangements or modifications of the output, but not the mere provision of prompts.
The Copyright Office confirms that the use of AI to assist in the process of creation or the inclusion of AI-generated material in a larger human-generated work does not bar copyrightability. It also finds that the case has not been made for changes to existing law to provide additional protection for AI-generated outputs.”
Unsurprisingly, the US Copyright Office reiterates human creativity’s centrality to copyright law and ties AI-enabled creativity to that principle. Also worth noting is the finding that there was no immediate need for new legislation to address copyright and AI. Obviously copyright law is different in different countries, but what the US does has sway elsewhere.
As Australia continues to grapple with this and other questions related to copyright and AI, it will be interesting to see if we take the same approach. One of the other questions we are looking at here in Australia is whether we should introduce an exception to allow the use of copyright material for text and data mining such as for training AI models. It is unlikely such an exception will get up, especially when the UK government’s proposed AI ‘opt out’ for their text and data mining scheme has seen criticism from music greats such as Sir Paul McCartney and Sir Elton John.
What’s worth reading on whether AI outputs should be protected by copyright:
Wednesday 29 January 2025
US Copyright Office
Carl Franzen – Wednesday 29 January 2025
VentureBeat
Laura Kuenssberg – Sunday 26 January 2025
BBC, British Broadcasting Corporation
Damian Jones – Sunday 26 January 2025
NME
A new copyright website-blocking bill has been put forward in the US
A new copyright website-blocking bill akin to SOPA/PIPA is being pushed for in the US.
US Congresswoman Zoe Lofgren has proposed the Foreign Anti-Digital Piracy Act, a copyright website-blocking law that would require ISPs (internet service providers) and DNS (Domain Name System) providers to block access to foreign websites facilitating piracy. Attempts to introduce a copyright website-blocking scheme in the United States have been made in the past. In 2010 the Combating Online Infringement and Counterfeits Act (COICA) failed to pass and a rewritten version called PROTECT IP Act (Preventing Real Online Threats to Economic Creativity and Theft of Intellectual Property Act, or PIPA), along with its sister Stop Online Piracy Act (SOPA), also failed to pass in 2012. You might remember Wikipedia, Google and other websites staged a blackout in protest. Representative Lofgren says the bill has come about “after working for over a year with the tech, film, and television industries” but players such as Public Knowledge and Re:Create have expressed their concerns.
Australia has a copyright website-blocking scheme that was introduced in 2015 and expanded in 2018. In a nutshell, section 115A of the Copyright Act allows copyright owners to seek injunctions requiring ISPs to block access to and search engines to not provide search results to an overseas internet location such as a website that “infringes, or facilitates an infringement, of the copyright” and “has the primary purpose or the primary effect of infringing, or facilitating an infringement, of copyright (whether or not in Australia).” Importantly, the scheme also establishes a ‘no fault’ enforcement regime through which copyright owners can extend the injunction to any “... domain names, URLs and IP addresses” the copyright owner and any ISPs or search engines involved agree in writing “… have started to provide access to the online location after the injunction is made.”
It is very encouraging to see that Representative Lofgren’s drafting includes consideration of a court that an injunction will not “interfere with user access to non-infringing material on another website or online service” and “disserve the public interest”. I wish we had a similar commitment to the public interest in the website-blocking scheme in Australia.
Despite a review of copyright enforcement in 2023 the extended Australian system still lacks public interest protections and allows too much potential for overreach and abuse of the scheme. As a former colleague of mine, Cheryl Foong, writing with Joanne Gray in the Australian Business Law Review back in 2020, noted, internet regulation is increasingly driven by private interests and Australia’s copyright website-blocking scheme is no exception. The public interest in the scheme is marred by a lack of explicit requirement for courts to consider the public interest in access to knowledge and innovation and because the regime pushes privately negotiated regulation (i.e. the parties towards private agreements, with limited transparency). As Foong and Gray put it:
“Unconstrained private ordering should not be considered a complete solution to online copyright enforcement because doing so risks setting up a legislative framework that leaves it to the largest industry stakeholders to determine a regulatory model that works best for them. Where there is a mix of private and public interests in a policy, law-makers should strive to retain some capacity for oversight and intervention to protect the public interest.”
What’s worth reading on the proposed Foreign Anti-Digital Piracy Act:
Wednesday 29 January 2025
United States Congresswoman Zoe Lofgren
Shiva Stella – Wednesday 29 January 2025
Public Knowledge
Wednesday 29 January 2025
Re:Create
Jon Brodkin – Thursday 30 January 2025
Ars Technica, Condé Nast
What’s worth reading on copyright website blocking in Australia:
Friday 16 December 2022
Australian Copyright Council
领英推荐
Cheryl Foong and Joanne Gray – 2020
Australian Business Law Review, Volume 48, Number 4, Thomson Reuters
Spotify thinks it is the saviour of music
Praise Spotify! Well at least if you believe the streaming giant’s recent rhetoric.
I haven’t seen anything more about Spotify’s mechanical royalties anomoly from last week, but the streaming giant has been very vocal of late about it’s role in the music industry and the amount it pays to copyright owners:
“In 2024, Spotify alone paid out a record $10 billion to the music industry—totaling nearly $60 billion since our founding.”
The post on Spotify’s corporate blog goes on to claim:
“... we estimate that, in 2014, around 10,000 artists generated at least $10,000 per year on Spotify. Today, well over 10,000 artists generate over $100,000 per year from Spotify alone. That’s a beautiful thing.”
Spotify were also spruking a new multi-year agreement penned between it and Universal Music Group that would bring the streaming platform and one of the ‘Big Three’ record labels even closer together. Despite being very vague, the media release does say that “Artists, songwriters and consumers will benefit from new and evolving offers, new paid subscription tiers, bundling of music and non-music content, and a richer audio and visual content catalog.”
But subscription bundling is what Digital Music News blamed for lowered mechanical royalties last week. Similar claims were reported in music news last year as well, with fears bundling audiobooks with music on Spotify would reduce royalty payments for recorded music by about $150 million. Just this week a US federal court found that Spotify was able to claim a reduced royalty rate because of bundled packages. The Mechanical Licensing Collective has not ruled out appealing the decision – and it’s very likely they will.
Still on Spotify, here at home, Ben Eltham took to the Nine newspapers to call out Spotify for it’s role in the decline of Australian music:
“Australian audiences have been captured by the global streaming giant. Big American pop stars rule the streams, and local music is suffering.”
Eltham says the steaming platform “serves predominantly international content to Australian ears”, and a lack of country-by-country break downs on playback on Spotify means the amount of Australian content streamed is not known publicly. We do know Spotify paid $275 million (AUD) in royalties in Australia in 2023. But, by Spotify’s own admission, “In 2023, more than 80% of all royalties generated by Australian artists on Spotify were from listeners outside of the country, with Brazil, Mexico, and Germany emerging as new export markets for local artists.”
As Eltham notes, there’s a bigger issue than Spotify’s impact on the Australian music scene: “You might be noticing a theme here. Local culture is being swamped by large overseas corporations.” Eltham reminds us that:
I agree with Eltham that more needs to be done to address the increased ownership of Australian culture by multinational corporations.
What’s worth reading on Spotify’s recent behaviour:
David Kaefer? – ?Tuesday 28 January 2025
For the Record, Spotify
Sunday 26 January 2025
Universal Music Group
Jem Aswad? – ?Wednesday 29 January 2025
The Verge
Jem Aswad? – Wednesday 29 January 2025
Variety
What’s worth reading on Spotify’s impact on the Australian music industry:
Ben Eltham? – ?Thursday 30 January 2025
The Sydney Morning Herald
Google’s renaming of the Gulf of Mexico is a reminder that mapping is political
Google’s decision to rename the Gulf of Mexico in the US is political, but maps have always been political.
Maps are political. Most people might think of them as objective representations of the physical world but they are full of socio-political meaning. There are lots of books on the topic (of which I own quite a few!) but Google’s decision this week to rename the Gulf of Mexico (in the USA at least) is an explicit reminder that cartography carries political meaning. In that assault of executive orders US President Donald Trump signed in the first 48 hours in office was an order changing the name of the Gulf of Mexico to the Gulf of America on US maps and in federal references.
Google took to X to announce they would change the name for US users. Google claims it has a longstanding practice of applying name changes made in official government sources, and when official names vary between countries Google Maps users see their local official name. Not long after Google reclassified the USA as a ‘sensitive country’, apparently “a designation it reserves for states with strict governments and border disputes”. I totally agree with CNBC’s Jennifer Elias that The decision reflects “the challenges Tech companies face in navigating the Trump presidency”.
The Mexican President Claudia Sheinbaum sent a letter to Google asking them to reconsider the renaming. As Sheinbaum noted, the United Nations Convention on the Law of the Sea dictates that an individual country’s sovereign territory only extends up to 12 nautical miles out from the coastline and so the US’s name change could only correspond to the 12 nautical miles away from its coasts.
Trump also ordered the name of Denali be changed to Mount McKinley. The move is equally political. Google will also change the name on Google Maps. America’s tallest mountain has been known as Denali to the Koyukon people who inhabit the area around the mountain for centuries. It was renamed Mount McKinley in 1896 and formally named such by the US federal government until 2015 when the official name was changed to Denali. On Friday 24 January 2025 the US Department of the Interior changed its name officially back to Mount McKinley. The whole affair is just another example of poor relations with First Nations communities, including Native Alaskans.
As an extra maps aside, I read this fascinating article about a hidden secret in a map that linked the British colonists of Roanoke Island with shards of British ceramics discovered in present-day Bertie County, North Carolina. This is a must-read for other map nerds.
What’s worth reading on Trump’s executive orders related to map locations:
Tuesday 28 January 2025
Reuters
Jennifer Elias – Tuesday 28 January 2025
CNBC
Friday 31 January 2025
Reuters
What’s worth reading on the Roanoke map discovery:
Monday 27 January 2025
New York Post
A bit on the side
Other tasty tidbits this week:
More to read
I use Are.na to gather and organise links about things I am interested in. If you are interested in any of the topics in this weekly roundup too, take a look at these channels:
If you are on Are.na too you can even add links to those channels.
Was this free blog post helpful?
If so, I encourage you to please show your support through a small contribution – it all helps me keep creating free arts marketing content.
Disclosure
Conflict of interest
I am the Copyright Officer (part-time) at the Australian Digital Alliance (ADA) and at the Australian Libraries and Archives Copyright Coalition (ALACC). The views expressed in this blog post are my own and do not express the views of the ADA or the ALACC.
AI use
This blog post was drafted using Google Docs. No part of the text of this blog post was generated using AI. The original text was not modified or improved using AI. No text suggested by AI was incorporated. If spelling or grammar corrections were suggested by AI they were accepted or rejected based on my discretion (however, sometimes spelling, grammar and corrections of typos may have occurred automatically in Google Docs).
The icon in the banner image (i.e. the first image at the top of the blog post) was generated by AI using Text to Vector Graphic (Beta) in Adobe Illustrator. Prompt: { ‘An outlined question mark and exclamation mark’ }.
Credits
Image: A colourful icon of a question mark and exclamation mark. The question mark is in two shades of pink and the exclamation mark is in two shades of blue. Both sit on an orange background. The icon is an adaptation of an vector graphic generated by Elliott Bledsoe using the AI tool Text to Vector Graphic (Beta) in Adobe Illustrator.
Provenance
This blog post was produced by Elliott Bledsoe from Agentry, an arts marketing micro-consultancy. It was first published on Sunday 2 February 2025. It has not been updated. This is version 1.0. Questions, comments and corrections are welcome – get in touch any time.
Reuse
Good ideas shouldn’t be kept to yourself. I believe in the power of open access to information and creativity and a thriving commons of shared knowledge and culture. That’s why this blog post is licensed for reuse under a Creative Commons licence.
Unless otherwise stated or indicated, this blog post – WTF now?!: Monday 27 January–Sunday 2 February 2025 – is licensed under the terms of a Creative Commons Attribution 4.0 International licence (CC BY 4.0). Please attribute Elliott Bledsoe as the original creator. View the full copyright licensing information for clarification.
Under the licence, you are free to copy, share and adapt this blog post, or any modified version you create from it, even commercially, as long as you give credit to Elliott Bledsoe as the original creator of it. So please make use of this blog post as you see fit.
Please note: Whether AI-generated outputs are protected by copyright remains contested. To the extent that copyright exists, if at all, in the icon I generated using AI or the banner image I compiled using that icon for this blog post (i.e. the first image at the top of the blog post), I also license it for reuse under the terms of the Creative Commons licence (CC BY 4.0).