With the U.S. Presidential election stirring worries among policymakers, businesses, and citizens about how generative AI will speed up the creation and propagation of disinformation, it becomes crucial to understand the existing laws to combat disinformation and the other laws that might obstruct this goal.
Enhancing social media engagement through buying fake followers or making false declarations about business successes might require politicians and candidates to exhibit caution. Such deceptions may at times be regarded as an unjust and duplicitous trade practice.
The broad powers of the Federal Trade Commission (FTC) extend to investigating businesses involved in unfair and deceptive trade practices, including misrepresentations under Section 5 of the FTC Act. Misleading public claims, such as those made by oil and gas companies about having sustainable products, have faced penalization from the FTC.
Reflecting on the potential of artificial intelligence in disinformation, FCC has banned AI-created voices for unsolicited robocalls citing the ability of Generative AI chatbots like OpenAI’s ChatGPT to rapidly generate content in response to prompts, creating possibly incorrect and potentially damaging information about individuals.
Generative AI technologies are built on extensive datasets, often extracted from the internet. These technologies provide responses based on word distributions in their training datasets without a real understanding of the truth or falsity of statements, which may be partially mitigated through reinforcement learning during the training phase. However, this can lead to biases in datasets, insufficient data or overfitting of models resulting in inaccurate responses or ‘hallucinations’.
Defamation laws can offer recourse to victims of disparaging statements made by AI models. An instance of this was seen when a Georgian radio host sued OpenAI for defamation caused by a false claim made by ChatGPT about embezzlement of funds.
The First Amendment guards free speech making it challenging to enact laws that demand social media companies to take down content perceived as disinformation. These challenges bubbled up during the Covid-19 pandemic when the Biden administration asked multiple social media companies to remove false information about Covid treatments and vaccine effects – a move now under scrutiny for potentially violating the First Amendment.
A series of children’s privacy and content safety laws passed by various states have also been met with injunctions on the grounds of interfering with the First Amendment. It remains unclear whether the First Amendment would cover generative AI content, and whether requests to social media companies to prevent the amplification of generative AI output would violate the First Amendment.
Generation of AI models trained on private data may result in possible privacy violations through processing of personal data. Several states allow claims for intrusions on privacy and violations of a person’s reasonable expectation of privacy. There is no federal legislation covering deepfakes and their proliferation, while a few states have proposed bills aimed at countering the effects of deepfakes in political discourse.
Celebrities in particular, often fall back on right of publicity protections when their pictures or voices are used for advertising without permission and compensation. The California Decency Act’s Section 230 provides businesses a ‘safe harbor’ from liability for content posted by its users. Despite opposition against its repealing as it may affect online free speech, attempts to repeal and revise this provision are likely to influence governments’ ability to request social media companies to remove generative AI content or hold them accountable for disinformation on their platforms.
The applicability of the Section 230 protections towards generative AI companies that aren’t hosts of third-party content remains unclear. Even Meta Platforms Inc. was accounted for housing discrimination stemming from its advertising algorithms and was not given the benefit of Section 230 as Meta was considered a co-developer of discriminatory ads. Some bills propose apparently to clearly exclude such AI vendors from Section 230 protections.
Addressing disinformation has always been a complex area, and these issues continue to be amplified by the onset of generative AI. Legal professionals would particularly benefit by staying aware of their rights and their potential recourse as they navigate these uncharted territories of the AI landscape.