A consortium of thirty-four states in the US has filed a lawsuit against Meta, alleging that Facebook and Instagram have been manipulating kids who use these platforms. Legal counsel from a number of states, including Louisiana, Virginia, New York, Ohio, South Dakota, and California, claim that Meta utilizes features like the “Like” button to encourage addictive behavior and harm children’s mental health. As per reports, the head AI scientist at Meta has stated that worries about the technology’s existential hazards are still ‘premature’. AI has previously been used by Meta to address concerns about safety and trust on its platforms. The government litigants, however, are continuing their legal action.
For each state included in the paper, the state lawyers are requesting varying sums of damages, restitution, and compensation; the amounts range from $5,000 to $25,000 per alleged event. Cointelegraph requested additional information from Meta, but as of the time of publication, no response has been received. In related news, the UK-based Internet Watch Foundation (IWF) has expressed alarm over the startling increase in the amount of child sexual abuse material (CSAM) produced by artificial intelligence. The IWF recently reported that in just one month, more than 20,254 AI-generated CSAM photos were found in a single dark web forum. They cautioned that this increase in unsettling content could eventually flood the internet.
The UK organisation proposed a multimodal approach and called for international cooperation to tackle the CSAM problem. This entails modifying current legislation, improving training for law enforcement personnel, and putting regulatory oversight for AI models into place. The IWF recommends that AI developers focus on eliminating child abuse material from their models, forbid using their AI to generate information about child abuse, and exclude related models. The development of generative AI picture generators has greatly enhanced the ability to create human reproductions that are lifelike. Tools that can produce realistic images include Midjourney, Runway, Stable Diffusion, and OpenAI‘s Dall-E platform.