Artificial Shackles: WGA Offers Members Up As Digital Slaves To AI Overlords

Imagine a thief holding a gun to your head and forcing you to pack up his truck with all your valuables. As he drives away, you and your family smile and wave goodbye. This unsettling image draws an eerie parallel to what awaits WGA members if they vote in favor of the agreement presented by the Writers Guild of America (WGA). Regrettably, this agreement with the Alliance of Motion Picture and Television Producers (AMPTP) offers scant protection against the looming challenges posed by AI.

This WGA agreement has sanctioned the sick and twisted practice of using a writer's past work to train the very AI models that will ultimately replace them. Due to ambiguous and non-committal language in the agreement, WGA members are afforded no protection from their past or future work being used for AI training. Without AI training transparency, writers will be left in the dark not knowing if their past or future work has indeed been used for AI training. The WGA also failed to set up a payment structure that would ensure compensation for the massive contributions its members have made to AI training. The technology cannot even exist without the work of WGA members. Thanks to this weak agreement, writers can't opt out of their work being included in AI training datasets that will be (and already have been) used to generate scripts which will one day soon replace them. Not preventing this perverse fate for its writers is downright reckless of the union. It is astounding that their negotiation committee presented this agreement to its members, knowing how eager they would be to get back to work. Many are faced with losing their homes or, sadly, not having the money to eat. This agreement's failure to shield its writers from such a grim fate is an irresponsible disregard for their welfare.

Perhaps one of the worst parts of this agreement is that it is a huge slap in the face of artists around the world. At this very moment, thousands of book authors and artists are in litigation fighting to have their intellectual property respected. They are fighting for fair compensation and the right to opt out of their life's work from being included in AI training datasets. Earlier this summer, WGA and SAG member Sarah Silverman filed a lawsuit against OpenAI and Meta for copyright infringement. Not only is her own writers union not protecting her with this agreement, it is actually undermining her lawsuit. By not acknowledging the past AI training contributions of its very own writers, which makes AI possible in the first place, the WGA has set a bad precedent within its agreement with the film studios. This oversight hurts artists globally who deserve fair compensation for the intellectual property AI utilizes.

By sanctioning the use of unethical AI in its films, the WGA has given the green light for all artists to be exploited by AI companies and studios worldwide. The precedent the union has set will surely be used to defend the usage of AI models in litigation. This ill conceived move may actually expose WGA members themselves to potential legal risks. The AI models are not only trained on all their past scripts but also the work of thousands of book authors. By allowing its members to work alongside unethically trained AI models to generate screenplays, the screenwriters may unknowingly plagiarize the work of others. The WGA agreement that it has presented to its members for a vote on Monday is totally irresponsible.

Other areas of the AI portion of the agreement are equally alarming. The lack of transparency creates a fertile ground for exploitation, as it fails to establish clear mechanisms for enforcing provisions or resolving disputes related to generative AI usage. This astounding lack of oversight may lead to legal ambiguity and conflicts between the WGA and studios. In one laughable stipulation by the AMPTP, they promise that “The Company shall disclose to that writer that the written material was produced by GAI (Generative AI).” In effect, the WGA is setting up an honor system with the studios that have been built on a foundation of exploiting artists and whose sole purpose has been to minimize profit sharing to stakeholders via the shady practice of "Hollywood Accounting." There are many such AI stipulations within the agreement that comically use this honor system. Without transparency and mechanisms such as audit rights, this agreement is worthless. It is also important to note that there is no effective way to detect AI-generated writing. As the technology progresses, it will be impossible to differentiate from human text. Without transparency, the WGA will have to take the film studios' word for it being legitimately written by a human.

The agreement's commitment to semi-annual reviews of AI is out of touch with the reality of the technology. With AI evolving at a blistering pace, semi-annual reviews are grossly insufficient. The rapid evolution of AI demands more frequent and dynamic assessments to adequately address and mitigate the unforeseen impacts and challenges emerging within short time frames. The WGA is ill-prepared for this technology that will sweep its members away in a relentless wave of advances. The saddest part is that this digital onslaught was foreseen and the risks could have been mitigated within this agreement.

Beyond the AI aspects, there is also reason for concern, especially when put through the prism of AI. In one area of the agreement that is being touted as part of this so called 'Exceptional Deal,' newly established writing staff minimums may not be all they seem. The minimums create an illusion of job security when, in fact, studios can now use AI for writing work not covered by the minimum staffing requirement. They can use AI tools for writing assistant roles that normally support writers and editors. They can use AI to generate outlines or scripts for lower level freelance assignments that would've normally been given to entry level writers, reducing opportunities for new WGA members. AI can still significantly displace non staff writing jobs without technically violating the agreement. This is a major oversight.

The agreement is being spun as a great deal, but that is far from the truth. Aside from the horrific AI portion of the agreement, most of the perceived gains are totally diminished by weakness in other areas. The foundation of the deal is built upon residuals. However, when you factor in growing inflation, these gains may actually equal a loss in the near future. There are also massive holes in the streaming portion of the agreement. In his LA Times op-ed, WGA member Nelson Chang pokes holes in the agreement, highlighting the murky language that “gave up on getting more money for writers from shows that perform well in exchange for minimum transparency from streaming services on viewership.” He reiterates that due to vague language in the agreement, the union is no longer using data on hours streamed to base residuals on as they originally proposed. They are "providing data on hours streamed and doing nothing with it." He also points out the bizarre clause that creates bonuses for projects but only for the ones that are viewed by a minimum of 20% of the streaming service's subscribers in the first 90 days of release. So essentially, this one time payment will not be honored for extremely popular shows that get millions of viewers and just don't pass this arbitrary threshold. The more you read this agreement and focus on the actual language it uses, the more problematic it becomes.

In the realm of AI, there is no room for ambiguity or grey area, especially in a paramount agreement such as this one. An agreement which is likely the most important agreement in the WGA's entire history. The implications of this agreement are massively profound and far reaching. If WGA members vote "yes" on this agreement, it will be the beginning of the end for human writers. The agreement must incorporate technical safeguards that preemptively address the complex interplay of AI and human creativity, with explicit clauses for ethical AI training, fair compensation structures, and protective measures safeguarding the creative and legal integrity of writers' contributions. The narrative is unfolding, and the final script is far from complete. The imperative now is for a technically informed, ethically sound, and legally robust agreement that not only foresees but adeptly navigates the intricate tapestry of challenges and opportunities presented by the advent of AI in content creation. The spotlight is on, and the onus is unequivocally on the WGA to script a future that is inclusive, fair, and visionary in its approach to integrating AI into the creative arts.