Summary
The Public Issues Committee for APANZ has sent an open letter to the Minister of Mental Health (Hon. Matt Doocey) outlining that Artificial Intelligent (AI) Social Systems are changing the relational foundations of our society. While AI companies release AI companions and chatbots for profit, they continue to do so unregulated and unaccountable for any harm or suffering encounter by the people (and their communities) interacting with these AI – especially in Aotearoa New Zealand. This political preference for a light-touch on AI regulation is grossly inadequate and we pushed for this to receive urgent attention.
Concerns and risks where highlighted which included; risks linked to AI Chatbot use, for example, psychological dependence, unhealthy emotional attachment, manipulations, misinformation, increased delusions and self-harm. This is especially concerning for children who interact with AI chatbots – whether online via smartphone, computers or even in toys. The AI chatbot harms that have already been emerging were also highlighted, which included suicides – where AI chatbots encouraged self-harm, harm to others and actively discouraged users help seeking; AI chatbots that can reinforce delusional thinking; AI chatbots claiming to be real people, and designing AI chatbots to keep users engaged and addicted.
Consequences were put forward, recommending regulators in Aotearoa New Zealand should be requiring AI companies to complete rigorous randomised controlled trials and safety testing, before allowing them to be released to the public. This would help the public to differentiate between scrupulous and unscrupulous AI companies and the chatbots and services they offer. It was recommended that there are a number of technical methods and model specifications that AI companies can implement to make an AI chatbot safer; and it should – especially when vulnerable people use it. These would include more interdisciplinary contributions to the development and uses of AI social systems and more education and transparency for the human end-user and third parties who purchase AI technology for human services.
Addressing the New Zealand government, we stated that as our society is becoming more and more reliant on AI technology, law makers and regulators who are in positions of power to make decisions about how AI should be regulated and controlled, should be urgently stepping in to do so. This includes 1) Restricting AI chatbots/companions for anyone under 18 years old, requiring real age verifications; 2) Requiring AI Social Systems to detect and interrupt human end-user emotional dependency, suicide ideation, self-harm and/or homicidal harm; 3) Requiring AI Social Systems to consistently and reliably redirect distressed human end-users to real-world human support and mental health resources; and 4) Establishing liability for harm caused, that hold AI companies accountable for psychological injury.
Currently, in Aotearoa New Zealand, there is no liability systems and accountability in place for companies whose AI Social Systems have/or will cause psychological and/or physical harm to a person(s). If the NZ government want AI that helps humanity flourish, start by protecting the New Zealand public – especially children. Put regulations in place to prevent unethical, unscrupulous AI systems from being unleashed to the public, in the first place, and then establish legislation that strengthens human connection, not replaces it.
Public Issues Committee for APANZ
