Less than a week ago, we witnessed one of the first whistleblowing incidents involving a big AI firm Open AI. This incident which the local police in San Francisco dubbed a suicide is not the end of a saga, despite the tragedy of the death of a young person, but the beginning of one, as the multiple lawsuits that may emerge from the details that the whistleblower, Suchir Balaji unearthed.
Whistleblowing has a long and hard tradition across the world and in particular, in the US. We discuss this issue in our Public Administration courses, not just the ethics class, but also in other core classes, including Public Admin Theory.
We have seen a host of high profile whistleblower cases : Snowden, Chelsea Manning and more recently Josh Paul, who protested against Biden government’s unqualified support for Israel, in its “war” on Gaza, which Amnesty International and even the UN has classified as a genocide. Josh Paul’s reason for resigning is that he the American efforts to support Israel without any regard for our own humanitarian laws and international norms as “shortsighted, destructive, unjust, and contradictory to the very values that we publicly espouse.
Each one of them can be seen as acting from a sense of higher self and a personal morality, even if it means breaking the law when the laws don’t uphold the morality of a society it is supposed to govern.
AI is an emerging field in some ways, however; the challenges that whistleblowing will throw up will be similar in nature to what has transpired in the past. The technicalities will involve well established field such as legal norms of copyrights, technology laws, among others.
There has been some debate on this issue, mainly from those studying and focusing on the ethical aspects of AI. As Timnit Gebru, a former Google computer scientists pointed out. She is a more recent whistleblower, who was fired from Google for pointing out the dangers of large scale language models. Her paper is what seems to have gotten her in trouble. The core of her thesis is that “facial recognition to be less accurate at identifying women and people of color, which means its use can end up discriminating against them.” Her argument was seen as challenging the status quo too much and in some ways subversive. The question that Gebru and her co-authors were asking was whether such large-scale language models were going to be used in ways that may harm some people and what steps google would take to prevent them. This seems to be a very relevant question in the context of AI and use of autonomous weapons in wars. We are already seeing this in the case of Ukraine and in Gaza. As a country that is one of the leaders in the field of AI, the US does have responsibility to regulate this space and make it safe for all Americans (and in some ways all human beings) as our technologies do impact people all around the world.
While the US and other countries have some form of whistleblower protection, there is much to be desired when it comes to protecting the whistleblowers. The gap between what is on the books and what happens is quite large. The value of media in exposing such unethical practices across the board is also crucial and at a time of increasing attacks on media, there is a greater need for vigilance on part of media.
As Ana Garrido Ramos speaks about her own experience as a whistleblower, the hardest part of the process is not the reporting the incident, but holding one’s ground and not giving in, to the pressure. Her tactics of appearing on media and exposing the corruption of local politicians in the greater Madrid area, where she worked as a government employee seems to have worked in her favor.
However, how are we to know that they are acting in everyone’s best interest and not just by their own moral compass (or desire for revenge)? This is hard one to answer and can only be answered with the passage of time.
Description:
Keywords. : AI, ethics, whistleblowing, government, copyrights