Safety & Harassment

The 2023 "Be The Change: Gender Equality in the Music Industry" study by Luminate, TuneCore, and Believe surveyed 1,656 music industry professionals and musicians from 109 countries. The study found that 34% of women in the music industry experienced sexual harassment or abuse at work, with rates rising to 42% for trans people and 43% for nonbinary people. Additionally, 58% of respondents disagreed with the statement that "everyone has an equal opportunity to succeed in the music industry." A separate 2018 Music Industry Research Association survey found that 72% of female musicians reported sex-based discrimination and 67% reported being victims of sexual harassment—nearly double the rate for U.S. women generally.

A 2024 Architects Registration Board (ARB) survey of 898 architecture professionals found that 41% had experienced bullying, 33% had experienced discrimination, and 10% had experienced sexual misconduct—rates higher than comparable professions including academia and medicine. Among female architects, 25% reported unwelcome sexual advances and 38% experienced unwelcome sexual comments, while 53% of women had encountered insults, stereotypes, or jokes relating to protected characteristics. A third of professionals said they would not feel confident raising concerns about misconduct, fearing career repercussions. A separate 2025 RIBA/Fawcett Society report found that 35% of women architects experienced sexual harassment at work, 54% discovered they were being paid less than male peers at the same level, and women in architecture represent only 27.1% of AIA membership despite decades of diversity initiatives.

A UNI Global Union survey of 28 unions in 22 countries, representing over 150,000 behind-the-scenes crew members, found that 62% said work schedule intensity negatively impacted their mental wellbeing. More than a quarter of respondents in independent television production reported that extreme fatigue had resulted in grave accidents. Crew members routinely work 14-18 hour days, with some working 16+ hours daily for 7 days a week. In 2014, crew member Gary Joe Tuck died after falling asleep at the wheel following an 18-hour shift on the set of "Longmire."

AI-generated deepfakes threaten performers' livelihoods and likeness rights. In 2024, Scarlett Johansson confronted OpenAI after it released an AI voice closely resembling hers without consent. In 2025, SAG-AFTRA filed an unfair labor practice charge over the AI-generated use of James Earl Jones' voice for Darth Vader in Fortnite, arguing it violated member rights and deprived living performers of potential work. The union's NO FAKES Act aims to create the first-ever federal intellectual property right in voice and likeness.

A 2024 European Labour Authority study found that 7.7 million workers in the EU's cultural and creative industries -- 3.8% of the total workforce -- face systematically precarious conditions. Reliance on self-employment and temporary contracts leaves the majority without essential protections including health insurance and pensions. Widespread undeclared work, compounded by cash payments and unregistered events, further erodes stability. A companion Creative Pulse Survey of 1,204 European artists found that nearly half reported poor working conditions, over two-thirds lacked sufficient social protection, and reliance on multiple income sources was most common in the performing arts. 96% of respondents supported adopting an EU-wide "Status of the Artist" framework to guarantee fair remuneration and social protections.

SAG-AFTRA's video game strike lasted from July 26, 2024 to June 11, 2025 -- nearly 11 months -- over AI protections for voice actors and motion capture performers. The core issue: companies sought the ability to train AI on performers' voices and likenesses without consent or fair compensation. The resolved contract was ratified with 95.04% approval, providing 15.17% compounded pay increases and AI consent/disclosure requirements. The strike highlighted systemic power imbalances between performers and studios.

Typeface designers face a uniquely difficult intellectual property landscape: their content is expensive to create but easy to steal. A study found that 63% of designers have used pirated fonts at some point in their careers. In the US, the shapes of typefaces are not eligible for copyright (only the font software is protected), leaving type designers with limited legal recourse. Notable lawsuits include Brand Design Co.'s $3.5 million suit against NBC Universal over approximately 20,000 unauthorized downloads of the Chalet typeface (2012), and Production Type's 2023 copyright suit against Nike for unlicensed use of Kreuz Light, seeking $150,000 per infringement.

A 2024 United Nations report condemned Japan's anime industry for exploiting workers, citing excessive hours, low pay, and disregard for intellectual property rights. The report warned of "potential collapse" if reforms are not implemented. Over one-third of anime studios reported losses in 2024 despite record revenues. In response, Japan enacted the Freelance Act in November 2024 — the country's first freelancer protection law — but enforcement remains uncertain for the estimated 50% of animators working as independent contractors.

Content creators, especially women, face escalating dangers from parasocial relationships that cross into obsession and violence. Twitch streamer Pokimane revealed years of stalking and sexual harassment by obsessive fans. Most women earning a living on Twitch know what it's like to have viewers develop obsessive feelings of romantic and sexual entitlement—extreme harassment, rape and death threats, blackmailing, and stalking have become regular workplace hazards. Platforms provide inadequate tools and no institutional support to protect creators from these dangers.

The xz Utils backdoor (CVE-2024-3094) revealed how a malicious actor exploited open source maintainer burnout through a three-year social engineering campaign. The sole maintainer, Lasse Collin, was pressured via sock-puppet accounts into ceding co-maintainer access after citing overwork and burnout. The attacker then injected a backdoor into a library used across virtually all Linux distributions. CISA concluded the incident "highlighted the fragility of key points in the open source ecosystem" and "the very real and ongoing risks created by maintainer burnout."

In August 2024, U.S. District Judge William Orrick ruled that visual artists may pursue copyright claims against Stability AI, Midjourney, DeviantArt, and Runway AI, finding that Stable Diffusion was built "to a significant extent on copyrighted works" and was "created to facilitate that infringement by design." The LAION training dataset contained 5 billion images scraped from the internet without consent. Over 30 copyright infringement lawsuits have been filed against generative AI developers, yet no final decisions have been reached, leaving AI artists in legal limbo -- unable to protect their work or establish clear rights.

A survey of over 700 international postdoctoral researchers at Harvard Medical School revealed that more than 40% spent over a month in their home countries waiting to renew US visas, with 5% spending more than six months -- effectively unable to work. Scholars from Asia experienced longer delays and higher costs than Europeans. In September 2025, the H-1B visa fee was raised to $100,000, while new requirements included mandatory disclosure of social media accounts and prolonged vetting. A Nature analysis found US-based scientists submitted 32% more applications to positions abroad in early 2025 compared to 2024, while applications to US jobs from European and Chinese scientists plummeted, signaling a brain drain from the world's largest research enterprise.

A Proof News investigation revealed that Nvidia scraped YouTube videos to train its Cosmos AI model, with VP of Research Ming-Yu Liu writing in an internal email about building "a video data factory that can yield a human lifetime visual experience worth of training data per day." Nvidia used dozens of virtual machines with rotating IP addresses to evade YouTube's detection systems. Companies including Meta, Microsoft, and Nvidia extracted over 15.8 million videos from more than 2 million YouTube channels without creator consent, prompting class-action lawsuits alleging unjust enrichment and unfair competition.

MIT Technology Review documented how AI companies have systematically scraped training data without consent or compensation, prompting a wave of lawsuits and licensing deals. The New York Times alleges millions of copyrighted articles were used to train AI models without consent. LinkedIn faces a class-action lawsuit for allegedly harvesting private messages for AI training. Reddit sued Perplexity AI for obtaining data "using false identities, proxies and other antisecurity techniques." Meanwhile, data scientists and researchers whose public datasets, code, and analyses were scraped for training have no mechanism for attribution, opt-out, or compensation -- their computational work treated as raw material for corporate AI products.

Open science hardware projects face a sustainability crisis mirroring the broader open-source funding gap: 60% of open-source maintainers remain unpaid for their work, according to the 2024 Tidelift State of the Open Source Maintainer Report. Many open science hardware projects emerge from academic labs or community initiatives but lack support to transition to self-sustaining operations. The Open Science Hardware Foundation was established to provide organisational support, fiscal sponsorship, and capacity building, but the fundamental tension persists: corporations and well-funded research institutions build on freely shared instrument designs while the original creators -- often early-career researchers or makers in the Global South -- capture no economic value. Open science hardware is generally perceived as lower quality than proprietary alternatives, creating a double bind where designers give away their work for free yet receive neither financial compensation nor professional recognition.

The 2025 tariff regime is reshaping the electronics supply chain in ways that disproportionately harm small and independent hardware makers. Russia, a key supplier of metals and minerals used in semiconductors, remains heavily sanctioned, and Ukraine's two largest semiconductor-grade neon producers -- responsible for 90% of U.S. neon supply -- shut down following the 2022 invasion, straining critical chip manufacturing inputs. Ongoing U.S.-China trade tensions have increased tariffs on semiconductor-related imports, reducing sourcing confidence and pipeline predictability. Small hardware engineers and scientific instrument builders lack the purchasing power to stockpile components or diversify suppliers across geographies. Unlike large manufacturers who can absorb tariff-driven cost increases, independent makers face existential risk: a single component becoming unavailable or doubling in price can render a custom instrument design financially unviable.

A peer-reviewed study in Anatomical Sciences Education evaluated AI image generators (Midjourney, DALL-E, Stable Diffusion) for anatomical illustration and found significant clinical inaccuracies: foramina were frequently omitted, suture lines were inaccurately represented, heart illustrations failed to indicate proper coronary artery origins, branching of the aorta and pulmonary trunk was often incorrect, and brain illustrations lacked accurate gyri and sulci depiction. Despite these dangerous flaws, AI generators produce images in seconds at near-zero cost, creating market pressure to replace trained medical illustrators whose work undergoes rigorous anatomical verification — raising the specter of clinically inaccurate visuals entering medical education and patient communications.

A 2025 peer-reviewed study in Anatomical Sciences Education documented growing concerns around using existing anatomical illustrations — including AI-generated derivatives — to produce new illustrations for academic publishing, raising issues of copyright infringement and plagiarism. The study noted that this problem has become more prevalent with the proliferation of publishing platforms and the increased adoption of generative AI in academia. For medical illustrators, this represents a double threat: their original works are scraped to train AI models, and the resulting AI outputs are then published as "new" illustrations that compete with — and potentially replace — the originals, all without attribution, consent, or compensation to the original creator.

The pharmaceutical industry funded approximately half of all U.S. continuing medical education costs, with industry support growing from $301 million to $1.2 billion between 1998 and 2007. The industry expects a return of $3.56 in increased sales for every dollar invested in CME, and 88% of clinicians surveyed acknowledged that commercial support introduces a greater risk of bias. Independent CME developers who refuse pharmaceutical sponsorship face a structural funding disadvantage, unable to compete with industry-backed programs that can offer lavish venues, free attendance, and extensive marketing. This creates a system where health educators must either compromise their independence or accept dramatically lower reach and compensation.

Unlicensed health influencers with millions of followers face lawsuits for deceptive marketing — Texas sued influencer Brittany Dawn Davis (1.3 million TikTok followers) for scamming customers with nutrition plans costing $92 to $300 that never delivered promised personalized coaching. Meanwhile, the FTC updated its Endorsement Guidelines to hold all parties in marketing campaigns liable for breaches, including content creators. Legitimate health educators face a double bind: they must comply with professional licensure, evidence standards, and liability frameworks that unlicensed wellness influencers ignore, yet the influencers' unregulated content attracts far larger audiences. Credentialed health educators who follow ethical guidelines are systematically outcompeted by those who don't.

More than 52% of companies targeted in patent lawsuits by non-practicing entities (NPEs) between 2017 and 2022 had annual revenues under $25 million, according to a study by HighTech-Solutions. The average cost to defend an NPE lawsuit hovers around $4 million, with patent trolls deliberately timing suits to coincide with funding rounds or acquisition events to maximize pressure for settlement. Small hardware firms shoulder a disproportionate burden, with patent abuse consuming approximately a quarter of their R&D spending. The number of small firms sued by patent trolls grew from 800 in 2005 to nearly 2,900 in 2011, with the median defendant's annual revenue just $10.3 million -- a scale typical of independent hardware makers.

Amazon removed over seven million fake goods from its platform in 2023, with estimates suggesting 10-60% of products in some categories may be counterfeit. For hardware makers, counterfeits create a dual threat: fake electronics contribute to over 70 deaths and 350,000 serious injuries annually in the United States, and when counterfeit versions of a creator's product cause harm, the original maker's reputation suffers. CBS News reported that $2 trillion in counterfeit products are sold globally each year. Independent hardware creators lack the legal resources to police counterfeits across global marketplaces, and Amazon's commingled inventory system means even buyers purchasing from legitimate sellers may receive counterfeit goods, eroding trust in the entire independent hardware ecosystem.

iFixit retroactively dropped the iPhone 14's repairability score after revealing that nearly half of iPhone parts are now "paired," meaning replacement parts require manufacturer-only proprietary software to function without triggering error messages or degraded performance. Even when a repairer installs a perfect, genuine OEM replacement part, the device will not work correctly without access to Apple's calibration or pairing software. iFixit applies a severe global scorecard penalty for parts pairing, noting that any repairability points gained through better physical design are negated if routine repairs cannot actually be completed outside of the manufacturer's authorized repair network. Oregon became the first US state to ban parts pairing in 2024, with its law taking effect January 1, 2025.

Art restorers and conservators face unique professional liability exposure: a single mistake on a valuable artwork can result in claims worth millions, as damaged pieces may suffer permanent diminution in value even after repair. Professional liability (Errors & Omissions) insurance is essential, covering situations where a client claims restoration damaged the original piece or lowered its value, with premiums reflecting the catastrophic financial risk. General liability, inland marine (covering objects in transit and in care), and professional liability policies must be layered to provide adequate coverage. For independent conservators and small restoration shops, these insurance costs represent a significant overhead burden that larger institutional conservation departments can absorb but solo practitioners struggle to afford, creating a barrier to independent practice.

The Gospel Coalition documents how sermon plagiarism has become widespread in evangelical churches, enabled by the internet's vast library of freely accessible sermon content. In a notable case, SBC president Ed Litton was found to have plagiarized sermons "nearly word for word" from another pastor for years, leading to the removal of over 140 sermons from his church's website. Commercial services now package and sell complete sermon content to pastors. The internet has acted as "kerosene on the fire" of pulpit plagiarism, creating an environment where original sermon writers see their creative intellectual work freely taken without attribution, while the theological and pastoral labor of crafting authentic sermons is systematically devalued.

Research examining translators in the platform economy found that work is predominantly outsourced by businesses in the Global North to workers in the Global South to reduce costs. In Turkey, where translators face record inflation, currency depreciation, and high unemployment, platform work has become one of few options to earn in foreign currency. Hungarian agencies pay translators approximately 50% less than their UK or Ireland counterparts, while Indian translators average just $0.04 per source word -- with minimums as low as $0.01 per word. The study documented substantial disparities in meeting fundamental conditions of decent work, including insufficient earnings, excessive working hours, difficulties achieving work-life balance, and limited social security access.

Academic research on crowdsourced mapping ethics found that "all is not well in the world of crowdsourcing at the moment, especially concerning fairness, respect and economical sustainability for the contributors." The study documents how crowdsourced mapping "might be losing its synonymity with communities of voluntary, often leisure-time mappers and the digital commons, becoming instead a new form of exploited labor," with individual mappers "reduced to a clickworker." Contributors often do not fully understand how their data will be used beyond the immediate mapping project, and their mapping data becomes permanently accessible to third parties. This dynamic depresses market rates for all cartographic work by creating a parallel supply of free, commercially exploitable map data.

Counterfeit board games have reached crisis levels, with 55.18% of consumers surveyed having unintentionally purchased counterfeit toys and board games online. Counterfeiters in China scan artwork, create component molds, and mass-produce copies using cheaper, untested materials. Some counterfeits originate from "third shift work" where factory workers produce bootleg copies at night using stolen materials. Small indie publishers with minimal components like Oink Games are prime targets because their games are easy to replicate at high margins. Counterfeit games sold through Amazon and other marketplaces directly cannibalise legitimate sales while the original designers receive nothing. Publishers must spend their own limited resources policing counterfeits rather than designing new games.

Children's toys rank among the top 10 most counterfeited product categories worldwide. Spin Master Corp. fought cases against Chinese companies engaged in "malicious infringement" of PAW Patrol and Bakugan IP rights, involving mass production of counterfeit and knockoff toys. Counterfeit toys bypass strict safety standards and regulations, posing direct risks to children. For independent toy inventors and puzzle creators, enforcing IP rights across international jurisdictions requires legal resources that dwarf their revenue. The combination of easy digital replication of designs, low-cost overseas manufacturing, and marketplace platforms that struggle to police listings creates a system where original creators bear all the risk of innovation while counterfeiters capture the profits.

CITES affords varying degrees of protection to over 35,000 species of animals and plants, and its regulations apply to any "readily recognizable part or derivative thereof," including taxidermy mounts. Taxidermists working with legally sourced specimens must navigate a labyrinth of permits for import, export, and interstate transport, with violations carrying severe criminal penalties. Moving any listed species across international borders requires CITES documentation even for personal or museum use. For taxidermists specializing in exotic or antique mounts, the compliance burden is substantial: pre-Convention specimen exemptions require provenance documentation that may be decades old, and regulatory uncertainty deters collectors from commissioning new work. This framework, while essential for conservation, creates significant legal exposure for legitimate practitioners.

A representative of The Union of British Columbia Performers, which represents 560 stunt performers, confirms a 100% injury rate among its members — catastrophic accidents, head injuries, substance abuse, anxiety, depression, and suicide have impacted every stunt performer in the community. Despite this, injuries routinely go unreported: performers fear being blacklisted, losing future work, or being perceived as unreliable. As one veteran performer put it, "A lot of injuries get covered up, quite frankly." The profession has no equivalent to the concussion protocols that protect athletes in professional sports.

Between 2002 and 2024, 152 catastrophic incidents — defined as fatalities or serious injuries — were reported on film and television sets, resulting in 32 deaths. The leading causes were falls (from ladders, catwalks, and rigging), followed by motor vehicle crashes during filming and stunt-related accidents. The deadliest year was 2011, with 5 deaths — 4 from falls. Workers categorized as "Laborers, except construction" were the most likely to be killed or injured, followed by "actors and directors." OSHA has documented over 40 stunt-related deaths in the U.S. film industry since 1980.

As Hollywood's content boom strains the ranks of seasoned stunt professionals, productions are hiring haphazardly and cutting safety corners. There are no SAG-AFTRA requirements stipulating qualifications for becoming a stunt coordinator — "If you have a SAG card, you can work as an actor, a stuntperson or a stunt coordinator." An official from SAG-AFTRA acknowledged that the jump in productions and their geographic dispersion creates "an increased risk of unqualified stunt coordinators" putting lives at risk. In one documented case, producers failed to place safety pads under sand where a performer crash-landed a Jet Ski at 28 mph, resulting in traumatic brain injury. Stunt performers report that SAG-AFTRA safety regulations have "no teeth."

On July 1, 2025, a pyrotechnics warehouse explosion in Esparto, California killed 7 workers and injured 2 others. Cal/OSHA issued 15 citations totaling $221,000 to Devastating Pyrotechnics, including failures to establish an Injury and Illness Prevention Program, train employees on emergency evacuation procedures, and train workers on fire hazards associated with explosive materials. The company also failed to immediately notify Cal/OSHA of the deaths. A subsequent State Fire Marshal investigation revealed "evidence of illegal activities," and the company's owner had previously been denied an ATF license following a criminal conviction — meaning he was legally ineligible to possess explosive materials.

WorkSafe Queensland issued a safety warning following the death of a stunt actor on an Australian film production, highlighting the lethal consequences of inadequate firearms and pyrotechnics safety protocols on set. In Australia, only accredited stunt performers graded by MEAA's National Stunt Committee across 6 skill categories (Body Control, Heights, Vehicles, Animals, Water, and Fire) may perform stunt actions — yet enforcement varies by state and territory. Research into Australian stunt performers revealed themes of "injury reporting stigma" and a "prevailing cowboy culture," with the Australian Stunts Organisation calling for specialized healthcare facilities and mental health support as performers struggle financially to maintain fitness and recover from injuries.