Briefing: Spotify says it will verify human artists (for now)

Our weekly Briefing is free, but you should upgrade to access all of our reporting, resources, and a monthly workshop.

Craig outlined 19 free tools you can use to investigate a URL. Learn tools and techniques for digging into links, redirects, parameters, and more.

Alexios wrote about viral “isolation challenge” videos on TikTok that enticed people with the (false) promise of a big payday if they could spend a few weeks in a luxurious-but-secluded location. The posts were lures for advance payment scams; TikTok removed hundreds of them following our outreach.

A nice thing: Indicator is nominated for General Excellence in Digital Publishing at the Digital Publishing Awards.

📍 Spotify announced it will add a badge to artists it has verified as human based on a range of signals, including “concert dates, merch, and linked social accounts.” The company wrote in the announcement that “profiles that appear to primarily represent AI-generated or AI-persona artists are not eligible for verification,” but premised this sentence with an “at launch” that suggests the policy may change over time…

📍 The 19th reports that Minnesota’s Senate unanimously approved a ban on AI nudifiers. The bill just needs Governor Tim Walz’s signature to become law.

📍 Taylor Swift filed trademark applications for two voice clips and an image of herself in a move that experts say is a precautionary measure against AI impersonation.

📍 The shooting at the White House Correspondents’ Dinner promptly spawned conspiracy theories claiming it was staged. Extra credit to the people who think a time-traveling AI warned that this was going to happen.

📍 South Africa’s Minister of Communications and Digital Technologies withdrew a proposed AI policy after it was found to contain likely AI-generated fabrications. He said that the errors “compromised the integrity and credibility of the draft policy.”

📍 Americans reported $2.1 billion in losses from social media scams in 2025, according to the latest data from the Federal Trade Commission.

📍 Three Arizona women have sued the men behind AI ModelForge, a network of AI porn influencers. The plaintiffs claim their likeness was used to generate the synthetic models, and that AI ModelForge actively trained other men to trawl the internet for photos of real women to use in their own nonconsensual pornographic schemes. Some of this material is apparently still live on Instagram and Fanvue.

We added over 2,000 new tools to OSINT Navigator.

When it launched a couple of weeks ago, Navigator suggested digital tools for your investigation based on a database of just over 7,500 entries drawn from 9 online OSINT toolkits. Now it includes over 9,700 tools from 17 lists.

The additional 8 sources are publicly maintained lists hosted on GitHub and elsewhere. You can view all 17 sources by clicking here:

Now you can go to Navigator, type in a search like “How do I investigate a crypto wallet,” and receive an improved list of potentially useful tools. Navigator offers 10 free searches a day, and shows which toolkit(s) suggested each tool. Tom Vaillant, the developer of Navigator, has also open sourced the data.

This is a great time to try Navigator if you haven’t already!

If you find it useful, you can upgrade to an Indicator membership and get 50 Navigator searches per day, MCP and API access, as well as access to all of our content and workshops. — Craig

📍 Aleksandra Bielska offered tips for how to access hidden results tabs in Google Search.

📍 Aidan Raney shared 13 OSINT tools you can use to investigate Reddit.

📍 Dmitry “Sox0j” Danilov made updates to Maigret, his open-source username enumeration tool. He said he rewrote the code and added “AI analytics and identity resolution, so you can get fast conclusions about a target across 3000 sites.”

📍 Mario Santella shared a link to ShareAll, which aggregates a wide range of searching engines, chatbots, and other tools.

📍 Nico Dekens wrote, “Vibe Coding Is Becoming an OSINT Risk."

📍 Andrew Deck of the Nieman Journalism Lab wrote, “Geospatial AI is reinventing the rainforest beat.”

📍 OSINT for Ukraine and The Hague Humanity Hub are offering a paid workshop on May 19, “Mapping Incidents for Accountability – A Practical Introduction.” It costs €35.

📍 Three researchers at the Oxford Internet Institute trained several LLMs to be warmer and more friendly in their responses. They found this also made the models far likelier to agree with inaccurate opinions. This could be a factor with AI companion bots that are typically developed to be support-providing rather than information-surfacing.

📍 The DFR Lab wrote about a pro-Kremlin effort to spread “unfounded EU economic loss figures in Estonia’s Russian-language information space ahead of October 2025 municipal elections.”

📍HKU journalism professor Masato Kajimoto shared a slide deck that maps out the tensions of legitimacy frameworks deployed by fact-checkers in different media ecosystems. In a comment, public health researcher Tina Purnat wrote that “if verification is always going to be read as agenda-driven, the more useful question is how to work within that reality -- not keep arguing against it.” As the person who helped bake “a commitment to nonpartisanship” into the fact-checkers’ code of principles, it’s haunting but crucial food for thought. — Alexios

📍 Reset Tech documented over 350,000 Meta ads (and more on Google) that often impersonated celebrities, doctors, and pharmaceutical companies to sell “unregulated and potentially dangerous nutraceutical products” in the EU.

📍 What To Fix found that Meta, TikTok, YouTube, and X are monetizing accounts that belong to EU-sanctioned individuals and entities.

Want more studies on digital deception? Paid subscribers get access to our Academic Library with 75 categorized and summarized studies:

Please upgrade to access to all of our content, including our how-to guides and Academic Library, and to our live monthly workshops.