Researcher beware: four red flags with the TikTok API's terms of service
TL;DR: Read the terms of service!
🚩 #1: Platform data retention policies make it difficult to do research and may be at odds with institutional data retention policies
🚩 #2: Required advanced notice of publication can harm independent research(ers)
🚩 #3: Your name and research is automatically licensed to TikTok in perpetuity
🚩 #4: Be aware of indemnity and forced arbitration clauses
Disclaimer: I am not a lawyer and this blog post does not constitute legal advice. I do, however, hope that you consult legal counsel before using the TikTok API.
Exit stage left: the demise of the Twitter Academic API
Twitter has long been viewed as the drosophila or HeLa of social media research. However, planned changes to its API will make it more expensive while simultaneously providing researchers with little to no data access — making public-interest research on Twitter difficult.
Enter: TikTok and its Research API
Seven months ago, TikTok first pledged to provide researchers with access to an API so that they can study the social media platform as well as its moderation policies. Recently, the company has started to accept applications to use their Research API from US-based academic researchers.
In Computational Research in the Post-API Age, Deen Freelon argues that researchers should understand the potential consequences of violating platforms’ terms of service (ToS). By reading through the ToS and speaking to other social media researchers, I identified four potential red flags that aspiring TikTok researchers should be aware of before signing up for the API.
Reasonable requirements, at first
The requirements to sign up seem reasonable at first: (1) having relevant and demonstrable academic research experience; (2) not having any conflicts of interest; (3) being employed at a US non-profit academic institution; (4) providing a clearly defined research proposal; and (5) only using data for non-commercial purposes.
However, API access is initially restricted to researchers at US institutions, with applications becoming open to those in the rest of the world at a later, unspecified date.
Requiring a research proposal is common — the Twitter Academic API and the YouTube Research API have similar requirements (Reddit’s API does not) — this also gives platforms the ability to decline proposals that aren’t in their best interest (h/t to Megan Brown at NYU CSMaP for pointing this out). Platforms can also block API access after the fact. For example, Facebook blocked Laura Edelson and Damon McCoy’s Ad Observer from accessing its API claiming that their tool violated Facebook’s terms of service. Yet, Edelson and colleague’s research using Ad Observer was invaluable in pointing out discrepancies in Facebook’s database of political ads. TikTok could similarly decline to provide access to researchers.
Additionally, the TikTok API terms of service (try saying that ten times fast!) has a data retention and refreshment policies, as well as a 15 day notice requirement prior to research being published. These policies also seem straightforward and easy to navigate, but in reality is far more complex.
It’s all fun and games until you really read the TikTok API terms of service
Taking a closer look at the terms of service (ToS), there are four red flags that researchers may want to watch out for.
🚩 #1: Platform data retention policies make it difficult to do research and may be at odds with institutional data retention policies
TikTok requires researchers to “regularly refresh TikTok Research API Data at least every fifteen (15) days, and delete data that is not available from the TikTok Research API at the time of each refresh.” This policy seems to be in line with: a) existing policies from other platforms (e.g., Twitter and YouTube), and b) ethical norms and participant expectations for handling data about people who aren't public figures.
TikTok’s requirement that researchers delete data that the API no longer returns is presumably because a user or TikTok has deleted or hidden the post. This can be problematic in two ways, however. Note that this limitation exists for Twitter as well, but not YouTube; Reddit does not clearly articulate their policy.
First, it makes it difficult to conduct research on social media data if the data set is constantly changing. In my own research studying election misinformation on TikTok, posts seemed to be taken down in a matter of hours or days, meaning I could no longer view them. While beneficial for the platform, it made it difficult for me to study TikTok. More importantly, it may be impossible to document human rights abuses or insurrections, or replicate research studies.
Second, TikTok’s (and Twitter’s) data retention policies may be at odds with one’s institutional data retention policies for research. Institutional data retention policies allow for research to be audited for research malpractice and harms to participants (often up to five years later). Which policy should a researcher follow? The one imposed by a multi-billion dollar corporation with an enormous legal team or the one imposed by your employer? If the underlying data for your research has to be deleted, do you need to retract publications based on that data?
Unlike TikTok and Twitter, the YouTube API allows researchers to cease data refreshing when one needs “the [data] to be fixed as to a point in time” to finalize data analysis and write up findings. Whether or not platforms with data retention and refreshment policies will actually enforce these policies is unclear. The question is if you, as a researcher, want to take that risk. (How much) will your institution support you? Do you have tenure or are you a first-year graduate student? Can you afford to hire a lawyer on your own if TikTok were to sue you?
🚩 #2: Required advanced notice of publication can harm independent research(ers)
TikTok requires researchers to provide it “with a copy at least thirty (30) days before its publication for courtesy.” YouTube requires seven days advanced notice, while Twitter and Reddit do not require any notice.
What’s the harm? Well… by providing TikTok with a copy of your research prior to publication, it might give them time to prepare for potentially negative news headlines. They might preemptively publish a blogpost discrediting your research, or discrediting you as a researcher. This could lead to irreparable reputational harm. While the chances of this happening are slim, they are possible. The tobacco, pharma, and fossil fuel industries have done this in the past, as have other social media companies.
More drastically, TikTok could choose to sue you in court, cite the potential defamatory nature of your work (building on your prior work), and request an injunction that prevents your research from ever being published.
The average researcher is unlikely to encounter these scenarios. But with greater research impact comes the possibility of greater pushback.
🚩#3: Your name and research is automatically licensed to TikTok in perpetuity
This one’s a doozy:
“You agree TikTok and its affiliates will have a worldwide, free, non-exclusive, perpetual or for the maximum term allowed by local law, sublicensable, unlimited license to use the Qualified Research Partner name and, logo(s), other identifying information, and to use, modify, reproduce, distribute, transmit, display, and create derivative works from your Research in whole or in part […] for marketing and promoting TikTok Research API and related services.”
I am not a lawyer, but I would interpret this as: “TikTok can use your name and other identifying information (such as your employer and their logo) as well as your research to use in perpetuity for any reason, including marketing TikTok’s API and the platform itself.” There seem to be two main problems with agreeing to this clause.
First, you probably, legally cannot hand TikTok a perpetual license to your published research and to your employer’s name and logo. This is because publishers often retain a license to publish your research (e.g., ACM) unless you choose to pay for open-access. Your employer also retains the rights to the use of their name and logo; you probably aren’t authorized to hand it over to TikTok.
Second, giving TikTok a perpetual license to your research would let them say on their website or in a Congressional hearing, something like “Our research partner, Jamila Doe at Great University, showed that TikTok is actually great for [insert societal issue here].” This could put you and your employer in the hot seat and result in a legal battle between you, your employer, and TikTok. You also probably don’t want TikTok referring to you as a “Research Partner” in advertisements or in Congress.
The ToS for YouTube, Reddit, and Twitter’s APIs do not seem to include similar licensing clauses. Reddit and Twitter also explicitly state that agreeing to the ToS does not constitute the formation of a partnership.
🚩 #4: Be aware of indemnity and forced arbitration clauses
Finally:
“To the maximum extent permitted by applicable law, you agree to hold harmless and indemnify TikTok […] from and against any third-party claim, suit or action including any liability, losses, damages (actual and/or consequential), expenses, litigation costs and reasonable attorney fees, of every kind and nature arising from or in any way related to your (a) actual or alleged breach of your obligations under these Research API Terms; (b) your use of Services or data derived therefrom; or (c) the Research.”
Again, I am not a lawyer, but to me this reads as “if a third-party sues TikTok based on research you conducted, you will have to indemnify TikTok,” i.e., compensate TikTok for any losses incurred. This, too, is possible but seems highly unlikely.
Interestingly, the TikTok, Reddit, and YouTube APIs’ terms of services did not have forced arbitration clauses, while the Twitter API ToS did. What this means is that you would be forced to settle a disagreement with Twitter through a third-party (i.e., you couldn’t take them to court).
What are your options?
Here’s what you can do:
🙅 Do nothing/wait. Don’t study TikTok. Don’t use their API.
🤔 Make an informed decision. Understand the risks. Consider speaking to your institution’s legal counsel. Use the API. Stay compliant with the ToS and keep an eye out for updates to it.
🤝 Find alternative solutions to using the API, such as scraping, data-sharing solutions (e.g., ICPSR), or third-party APIs and tools (e.g., Junkipedia, SMAT, or Communalytic). Note that some of these options may violate TikTok’s ToS, but may still be legal.
I’m pushing forward with both options 2 and 3; something about eggs and baskets and whatnot.
Regardless of which option you choose, if your research is in the public interest, consider advocating for greater access to social media data in public and with your elected representatives. As Megan Brown said, “platforms should not be the arbiters of what data is available, who gets access to data, and what data retention policies are required for access.”
If you found this blog post helpful, share it with a friend (or ten 🙃) who might benefit from reading it.
More: action, resources, ideas
✍️ Psst. Hey, you! Yes, you. Are you also worried about the Twitter API going away or becoming too expensive to use? Consider signing this open letter from the Coalition for Independent Technology Research calling on Twitter and policymakers to ensure continued access to Twitter’s API. I’ve signed it.
🎙️ Ethan Zuckerman interviewed Brandon Silverman, the co-founder of CrowdTangle, about the Platform Accountability and Transparency Act that Brandon is advocating for in Congress. You can listen to it here, or on your favorite podcast app.
💡 If you’re really really interested in platform policy — particularly API policy and how it governs data access — you might want to do a more systematic, comparative analysis of their ToS (similar to this one). I hear the Journal of Online Trust and Safety is a great venue to publish this type of research.
To cite this article:
Sukrit Venkatagiri. February 23, 2023. Researcher Beware: Four Red Flags with the TikTok API's Terms of Service. Technomoral. https://technomoral.substack.com/p/researcher-beware-four-red-flags