BOCHUM, Germany — AI technology presents many ethical and regulatory challenges. The media industry, in particular, is grappling with complex issues related to the multifaceted ethical and legal implications of AI-produced generative content, complicated by the unprecedented speed of progress in the field – beyond capacity regulatory bodies to follow.
In February, OpenAI announced Sora, a generative AI capable of producing high-quality video clips based on text prompts. This followed in March with Voice Engine, which can convincingly recreate an individual’s speech patterns from a 15-second clip. Although neither product has been released to market and the company is working with the developer community to establish guardrails around their use, their very existence has raised a range of issues.
These include, for example, ethical questions linked to the replacement of human speech. In radio production alone, there are multiple use cases, including reading news reports in the middle of the night, automating links between music tracks by reading the appropriate metadata fields, and targeting advertising at the individual level of the streaming listener. We can consider some AI improvements, but others, like automatic graveyard change coverage, would definitely replace a human.
Widespread concern
Although such use benefits a small radio station that can do more with less, corporate CEOs have shown much enthusiasm for downsizing, with all the ethical considerations that come with it. And that’s just the beginning of the puzzles. If on-air talent is replicated using AI, are the appropriate contracts in place? Who owns the voice? Has permission been granted for its use, and under what circumstances? If the talent leaves the station, do they take their intellectual property with them, or do those rights belong to the station in perpetuity? If so, what is fair compensation? Were the AI “clones” trained on closed or open data from the Internet? In the latter case, their production could constitute a violation of the law.
These are all profoundly difficult questions with far-reaching consequences. There is also widespread concern about this technology being used by bad actors. Individual countries are developing legal frameworks to cover these and other situations, but consensus is lacking at the international level.
Many initiatives integrate watermarking technologies into human- or AI-generated content so users can tell if the content is AI-generated. So far, no universal solution has been found.
The law on artificial intelligence
One of the challenges facing regulators and developers is the speed of progress and the seismic nature of some of its implications. Even the broadcast industry, a sector accustomed to rapid change, faces constant evolution. Governments are slower to act. The EU is leading the regulatory charge in the same way it did by introducing GDPR to prohibit limits on the use of consumer data. The Artificial Intelligence Act was passed on March 13 and became law on August 2.
It subdivides the use of AI into high, limited and minimal risks. Use in the creative and media industries will likely fall into the latter two. The law will impose conditions on media outlets regarding the ethical use of intellectual property and the transparency of the models that create them.
Where AI is driving content, transparency will prove essential.
Once the European Council adopts the law, its application will begin within six to 24 months to give the industry time to comply.
This is of course a geological era in relation to the pace of development of AI. Meta’s large Llama2-70B language model took a month to train when it was released in July 2023. It is estimated that this could be achieved in a single day using the latest chips. One of the key developments this year will be on-device AI, which will allow users to securely access generative AI models without going through the cloud.
Transparency is essential
So how can broadcasters navigate these choppy and ever-changing waters? One of the lessons learned from the introduction of GDPR is the vital importance of trust. This must permeate everything from interactions with end users to the collaboration networks that modern broadcasters and their suppliers build in the broader broadcast ecosystem. When we use generative AI or when it is used in our products by a plugin, for example, this use must be reported. It must be clear to the listener that the voice coming from their radio at 3 a.m. is not that of a real human being and, similarly, metadata reporting AI generative content must follow this content throughout a newsroom system.
This does not have to be the case for all AI use cases. Many everyday tasks have already shifted to AI, from cleaning up dialogue on a noisy recording to translating text. Yet where AI is driving content, such transparency will prove essential.
Regulation and public perception regarding the use of generative AI are currently behind the development curve. By being open and transparent in their use and ensuring that their use follows defined ethical guidelines, broadcasters can ensure they are ahead of the coming changes.
The author is vice-president of consulting services at CGI.
More intelligence, less artificial
Benztown introduces AI Listener voice generator
The Innovators: AI Voice Technology for Broadcast: A Professional Audio Perspective
