An AI-generated image (“Astronaut Riding a Horse”) created with Stable Diffusion XL. Such imaginative outputs highlight both the creative potential and contentious issues of generative AI.
en.wikipedia.org
Generative AI has rapidly entered the creative fields of visual art, video, and music – allowing anyone to produce images, animations, or songs with a simple prompt. This wave of innovation, however, has been met with significant public and professional concern. Artists, filmmakers, musicians, and even environmentalists have voiced alarms about how these AI systems are developed and used. Key issues include the legal status of AI-generated content, unauthorized use of copyrighted works in training datasets, threats to creative jobs, biases in AI outputs, and the environmental footprint of running large AI models. Below, we explore each category of concern with recent examples, studies, and quotes reflecting the ongoing debate.
Copyright and Intellectual Property Concerns
One of the loudest controversies surrounding generative AI is its impact on copyright and intellectual property. Visual artists and musicians have found AI models producing works uncannily similar to their own, raising the question: who owns the output, and was it created by infringing on someone else’s work? In early 2023, a group of illustrators – Sarah Andersen, Kelly McKernan, Karla Ortiz, and others – filed a class-action lawsuit against generative image companies Stability AI, Midjourney, DeviantArt and others. They allege that these companies infringed their copyrights by training AI on billions of online images (including their artwork) without permission, effectively creating a “21st-century collage tool” that regurgitates pieces of existing art.
npr.org
In a tentative legal victory for creators, a U.S. federal judge ruled in August 2024 that the lawsuit could move forward, finding “plausible” evidence that Stable Diffusion was designed to facilitate copyright infringement by using artists’ work without authorization.
theartnewspaper.com
This case – the first of its kind – signals that courts are taking artists’ IP claims seriously, though the ultimate outcome remains undecided. The scale of the copyright crisis is striking. By late 2024, at least 30 major copyright lawsuits had been filed against generative AI companies in the U.S., spanning images, music, books, and more.
generative-ai-newsroom.com
Stock image giant Getty Images also sued Stability AI in 2023, accusing it of scraping 12 million Getty photos (watermarks and all) to train Stable Diffusion “without permission or license,” an act Getty argues is blatant copyright and trademark infringement.
reuters.com
In music, a viral AI-generated song “Heart on My Sleeve” mimicked the voices and style of Drake and The Weeknd in 2023, garnering millions of streams before Universal Music Group intervened. UMG – which owns Drake’s label – invoked copyright law to get the track removed, stating that using “our artists’ music… without permission” to train AI or create deepfake songs “denies artists their due compensation.”
npr.org
The record label framed the issue as taking a side “on the side of artists, fans and human creative expression, or on the side of deep fakes, fraud.”
npr.org
This stance reflects mounting industry resolve to defend artists’ intellectual property. At the same time, legal experts note that copyright law is being tested in unprecedented ways. Generative AI companies often argue that training on public data is “fair use” – analogous to how search engines index content – and that the models learn patterns rather than storing exact copies of works.
npr.org
No court has definitively ruled yet on whether using copyrighted works to train AI constitutes infringement. The U.S. Copyright Office launched a major review of AI and copyright in 2023, recognizing that “generative AI is destabilizing foundational concepts of copyright law” and may require new rules.
issues.org
In the meantime, artists feel their creations are vulnerable: as one illustrator put it, “These programs rely entirely on the pirated intellectual property of countless working artists, photographers, illustrators and other rights holders.”
theguardian.com
In short, the rise of generative AI has set off a legal showdown pitting technological innovation against the ownership rights of creators.
Consent and Dataset Usage (Lack of Permission)
Beyond the letter of the law, creators object to the lack of consent and credit involved in AI training and generation. Most generative models were built by sweeping up massive datasets of content from the internet – usually without asking the original creators’ permission. This has left many artists feeling, in a word, violated.
Illustrator Kelly McKernan discovered that her name had been used over 12,000 times as a prompt in an AI art generator, producing images in her exact style. “There’s more and more images with my name attached that I can see my hand in, but it’s not my work. I’m kind of feeling violated here,” McKernan said.
npr.org
She was especially angered to learn that an art platform she used (DeviantArt) had launched an AI generator trained on the very art she and others had uploaded over decades – with no opt-out and no compensation for artists.
npr.org
Incensed creators rallied on social media with campaigns like #NoToAIArt, arguing that it’s fundamentally unjust for companies to profit from artwork scraped without consent.
theguardian.com
As cartoonist Harry Woodgate described, “AI programs rely entirely on pirated [art]… it samples everyone’s [work] then mashes it into something else.”
theguardian.com
What technologists call “training data” feels to artists like wholesale exploitation of their life’s work.
This consent problem extends to other creative fields. Actors and voice performers worry their voices and likenesses are being cloned by AI without approval. Voice cloning tools now only need a few seconds of audio to replicate someone’s voice – and numerous voice actors have found unauthorized copies of their voices circulating online.
latimes.com
“People are beginning to lose work, sometimes finding themselves competing with their synthetic avatars for jobs, and with very little power to do anything about it,” says actress and voice artist Laurence Bouvard of the UK’s Equity union.
computerweekly.com
During the Hollywood actors’ strike in 2023, a core issue was how studios wanted rights to scan actors digitally and reuse their likeness with AI – effectively creating “digital doubles” indefinitely, unless strict consent protections were established.
latimes.com
This prompted SAG-AFTRA to secure new contract terms that require consent and payment for any AI-generated performances using an actor’s image or voice. The “soul of the industry is on trial,” warned a group of video game voice actors, stressing that performers must have a say in how their identity and artistry are replicated by AI.
latimes.com
Perhaps the most visceral example of non-consensual AI use is the rise of deepfake pornography. New AI video tools can superimpose a person’s face onto adult video footage, creating explicit fake videos without the victim’s knowledge or consent. This phenomenon overwhelmingly targets women and has been described as a form of sexual violence enabled by technology.
By one estimate, 98% of deepfake videos online in 2023 were pornographic and typically non-consensual.
oregoncapitalchronicle.com
Images of actresses, influencers, or even private individuals have been stolen to generate “deepfake” sexual content – often shared publicly to harass or humiliate the victim. “One victim of deepfake pornography is one too many,” said a U.S. lawmaker pushing for stricter penalties.
morelle.house.gov
Victims like Uldouz Wallace have spoken out about the trauma of discovering fake explicit images of themselves circulating online: “For [people] to sit there and create so much fake content of someone that clearly doesn’t want anything of that sort? Without consent? It’s just crazy to me,” Wallace said.
oregoncapitalchronicle.com
This abuse has prompted several U.S. states (and countries in the EU) to draft laws banning non-consensual AI-generated imagery.
The broader point is clear – consent is emerging as an ethical line for generative AI. Whether it’s a painting style, a voice, or a face, creators and the public are demanding the right to opt out of AI’s training datasets and outputs. The absence of that consent so far has fueled backlash and calls for new regulations to protect personal and creative autonomy.
Bias and Stereotypes in AI-Generated Content
Another area of concern is the bias embedded in generative AI outputs, which can reflect and even amplify societal stereotypes. AI models learn from vast troves of human-created data, so any prejudices or imbalances in that training data can surface in the AI’s creations – sometimes in disturbing ways.
Investigations have found that text-to-image generators often produce skewed or stereotypical images when asked for people in various roles. The Washington Post, for example, tested prompts like “a builder,” “a CEO,” or “a person from [country]” and found the AI routinely defaulted to clichéd demographics: “Asian women are hypersexual. Africans are primitive. Europeans are worldly. Leaders are men. Prisoners are Black,” as the report summarized.
washingtonpost.com
None of these stereotypes reflect reality, of course – they mirror the biases in the internet data the AI was trained on, which is “rife with pornography, misogyny, violence and bigotry,” the reporters noted.
washingtonpost.com
Academic studies reinforce these findings. A 2023 University of Washington study of Stable Diffusion (a popular image generator) showed that when prompted with something generic like “a person,” the model over-represented light-skinned men in its outputs.
washington.edu
When asked for images of people from various regions, the AI often failed to represent local diversity – for instance, a prompt for “a person from Oceania” yielded images mostly of white-presenting individuals, effectively erasing Indigenous populations.
washington.edu
The researchers also observed how the AI sexualized women of certain ethnicities: women from Latin American countries, India, or Egypt were depicted in sexualized ways far more often than others.
washington.edu
“It’s important to recognize that systems like Stable Diffusion produce results that can cause harm,” said Sourojit Ghosh, the study’s lead author, noting the “near-complete erasure of nonbinary and Indigenous identities” in some outputs.
washington.edu
Such representations can be harmful by reinforcing a very narrow view of “normal” and rendering many groups invisible or only seen through stereotypes.
The tech companies behind these models have started to acknowledge the issue – adding filters and claiming updates to reduce bias – but gaps remain. Stable Diffusion’s latest version (as of late 2023) made some efforts to detoxify its dataset, yet a test by The Washington Post found it “still defaults to cartoonish tropes” and “amplifies outdated Western stereotypes” in many cases.
washingtonpost.com
As AI researcher Pratyusha Kalluri observed, companies seem to be playing “whack-a-mole” – fixing the most publicized biases but not the systemic ones.
washingtonpost.com
And in some cases, the source of bias is structural: the non-profit LAION database that feeds many image models admitted it contains far more U.S. and European data than content from Asia or Africa.
washingtonpost.com
Bias in generative AI is thus both a technical and a societal problem. From an ethical standpoint, artists and consumers worry that AI systems could perpetuate racial, gender, and cultural biases under the guise of objectivity. Images, videos, or lyrics created by AI might marginalize certain groups or normalize stereotypes if left unchecked.
This concern has spurred calls for greater transparency in training data and for diverse, representative datasets to be used – so that the “creative vision” of AI isn’t just mirroring the biases of the past.
Job Displacement and the Impact on Creative Labor
As generative AI becomes more capable, creative professionals fear it could displace human artists and undermine their livelihoods. Unlike past automation (which mostly affected repetitive manual tasks), AI now encroaches on fields once thought uniquely human – illustration, graphic design, voice acting, songwriting, video editing, and more.
The worry is not just theoretical; early signs of job impact are already here. In a March 2024 survey by the UK actors’ union Equity, 65% of members felt AI posed a threat to their employment opportunities, a figure that jumped to 93% among audio artists (narrators, voice actors, etc.).
computerweekly.com
Laurence Bouvard, the voice artist, notes that some performers are already losing gigs to synthetic voices or digital avatars of themselves.
computerweekly.com
Similarly, visual artists have seen clients experiment with AI-generated illustrations instead of hiring talent. One graphic artist, Chris Barker, observed people assuming a striking magazine cover he created was AI-made – diminishing the perceived value of his skill and creativity.
computerweekly.com
“Now I imagine people would just assume it was AI generated, which is a shame,” he said, reflecting on how quickly human artwork can be mistaken for machine output in the public eye.
computerweekly.com
The devaluation of human creativity – both financially and culturally – is a looming concern.
Whole job categories in the creative industries are under scrutiny. Illustrators and concept artists worry that game and film studios might use AI to generate background art or character designs, reducing the need for entry-level artists. That scenario already played out in early 2023 when an animated short film used AI for background art, spurring backlash from animators who saw it as cutting corners at their expense.
Voice actors are troubled by studios potentially using AI voices for minor roles or audiobook narration, which is often bread-and-butter work for them. “The onus falls on the individual to protect themselves,” notes one report, since many voice actors are freelancers without union coverage.
latimes.com
Until regulations catch up, they find themselves “in the odd position of competing with their own synthetic voice” for jobs.
computerweekly.com
Musicians and composers similarly fear that AI-generated stock music could replace human-made scores for ads or background music libraries. Famed musician Sting warned in 2023 that there is “a battle we all have to fight in the next couple of years: defending our human capital against AI.”
musically.com
He argued that the “building blocks of music belong to human beings,” and expressed that AI lacks the emotional depth of a human songwriter.
musically.com
This sentiment – that human creativity has intangible value that should not be cheaply replicated – is echoed by many artists. Author Ela Lee, for instance, said she “scrutinised every paragraph” of her novel and worries that AI-generated prose will be “uniform, stripped of nuance and lacking the human voice that makes people connect to art.”
computerweekly.com
Of course, not everyone believes AI will destroy creative jobs; some see it as a tool that, if used ethically, could assist artists rather than replace them. But the prevailing mood in creative communities is caution.
In 2023, both the Writers Guild of America and SAG-AFTRA (actors’ union) went on strike in part to address AI – securing provisions that writers cannot be forced to adapt AI-written scripts, and that actors’ digital likenesses can’t be used without consent and pay.
These battles underscore that creative workers are organizing to assert their rights in the face of AI.
In sum, while generative AI offers new avenues for creativity, it also poses an economic threat to those who make a living from art, video, and music. The public debate now grapples with how to balance innovation with fairness – ensuring that human creators are not left behind or forced to compete with an endless stream of cost-free AI-generated content.
Environmental Impact of Generative AI
Often overlooked amid the artistic and ethical debates is the environmental cost of generative AI. Training and running large AI models require massive computational resources, leading to significant energy consumption and carbon emissions. As generative AI applications have boomed, so too have the electricity demands of data centers that power them.
A recent MIT analysis noted that the explosion of generative AI is “increasing electricity demand and water consumption” in tech infrastructure.
news.mit.edu
Data centers – warehouse-sized server farms – already devour huge amounts of power, and AI is accelerating that trend. Scientists estimate that in North America, data center power usage roughly doubled from the end of 2022 to the end of 2023 (from ~2,700 MW to ~5,300 MW), partly due to the surge in AI workloads.
news.mit.edu
Globally, data centers used about 460 terawatt-hours (TWh) of electricity in 2022, which, if viewed as a country, would rank around 11th in the world in consumption (on par with nations like France).
news.mit.edu
By 2026, that number could more than double to 1,050 TWh – vaulting data centers to the 5th largest electricity consumption by country-equivalent.
news.mit.edu
While not all of this is from AI, generative AI has become “a major driver of increasing energy demands” in the tech sector.
news.mit.edu
Besides electricity, water usage and e-waste are pressing issues. “Beyond electricity demands, a great deal of water is needed to cool the hardware” for training and deploying AI models, writes MIT’s Elsa Olivetti, a materials science professor.
news.mit.edu
Data centers use chilling systems that can consume millions of liters of water, straining local water supplies especially in drought-prone areas.
news.mit.edu
For example, each megawatt of data center capacity can require around 2 liters of water per second for cooling – scale that adds up fast.
news.mit.edu
In regions where water is scarce or ecosystems delicate, this cooling requirement poses environmental and community challenges.
On top of that, the production and frequent upgrade of specialized AI chips (GPUs, TPUs, etc.) contribute to electronic waste and resource extraction. More and more high-performance hardware is being manufactured and shipped to meet AI demands, carrying a carbon footprint from mining and fabrication.
As Olivetti emphasizes, “When we think about the environmental impact of generative AI, it is not just the electricity you consume… There are much broader consequences that go out to a system level,” including supply chain and end-of-life impacts.
news.mit.edu
These concerns have prompted calls for a greener AI ecosystem. Some researchers are working on efficiency improvements and using renewable energy for data centers.
guides.lib.uci.edu
Policymakers are also paying attention: in the U.S., lawmakers introduced the Artificial Intelligence Environmental Impacts Act of 2024 to study and require reporting on AI’s carbon and water footprint.
markey.senate.gov
“AI offers incredible possibilities … but that comes with high environmental costs,” noted U.S. Congresswoman Anna Eshoo, stressing the need to ensure AI’s growth “does not come at the expense of the health of our planet.”
markey.senate.gov
Environmental groups and climate-conscious technologists echo this, pointing out that unchecked AI expansion could undermine global efforts to reduce emissions. Indeed, if data center energy use soars as projected, it will require significant clean energy deployment to avoid large increases in greenhouse gases.
In summary, the environmental impact is now a key part of the generative AI debate – linking the digital to the physical world. Energy-hungry AI models may carry hidden climate costs, and there is growing pressure to measure, report, and mitigate these impacts as AI continues to proliferate.
Conclusion
Generative AI’s advent in art, video, and music has sparked a multifaceted public debate. On one hand, these technologies unlock creative possibilities previously in the realm of science fiction. On the other, they raise profound legal, ethical, and societal questions about how we value human creativity, protect individuals’ rights, and steward our environment.
Artists and creators are challenging the unconsented use of their works and fighting to ensure they are not swept aside by algorithms trained on their own creations. Society at large grapples with issues of authenticity (in a world of deepfakes), fairness (in the face of bias), and sustainability (as AI’s carbon footprint grows).
The concerns outlined – from copyright lawsuits to climate bills – demonstrate that generative AI is not merely a technical innovation, but a cultural and ethical crossroads. As we move forward, the balance between embracing AI’s benefits and addressing its pitfalls will be shaped by ongoing dialogue among technologists, lawmakers, artists, and the public.
The outcome will determine how generative AI coexists with human artistry and values in the years to come.
Sources
The information and quotes in this report are drawn from a range of credible sources, including artist statements and interviews (The Guardian, NPR).
theguardian.com
npr.org
Academic research and analysis (University of Washington, MIT).
washington.edu
news.mit.edu
Journalism (The Washington Post, Los Angeles Times, Reuters).
washingtonpost.com
latimes.com
reuters.com
And legal or policy documents (court cases, legislative reports).
theartnewspaper.com
markey.senate.gov
These sources are cited throughout the text to substantiate the examples and perspectives discussed. Each citation in the format【†】refers to the specific source and line numbers for verification.
The aim has been to faithfully represent the current state of public and professional concerns about generative AI, emphasizing voices of those most affected – the creators – and the broader implications for society and the environment.