Exclusive: Google cancels AI ethics boards in response to rebellion

This week, Vox and other outlets reported that Google's innovative AI ethics boards fell apart in controversy over several board members.

Well, it is officially finished falling apart – it has been canceled. Google told Vox on Thursday that it pulls the plug on the ethics tray.

The board survived in barely more than a week. Founded to lead the "responsible development of AI" on Google, it would have had eight members and met four times in 2019 to assess concerns about Google's AI program. These concerns include how AI can enable authoritarian states, how AI algorithms produce different outcomes, about working with military applications of AI and more. But it went into trouble from the start.

Thousands of Google employees signed a petition requesting the removal of a board member, Heritage Foundation President Kay Coles James, about her remarks about the transsexuals and the organization's skepticism about climate change. [1[ads1]9659005] At the same time, the inclusion of Drone's CEO Dyan Gibbens opened the old divisions of the company over the use of the company's AI for military applications.

Board member Alessandro Acquisti submitted . Another member, Joanna Bryson, who defends her decision not to say, claimed James "Believe it or not, I know worse about one of the others." Other board members found swamped demanding that they justify their decision to remain on the board.

On Thursday afternoon, a Google spokesman told Vox that the company has decided to dissolve the panel, called ATEAC, with externally technical advisory board. Here is the company's statement in its entirety:

It has become clear that ATEAC in the current environment cannot function as desired. So we end the council and go back to the drawing board. We will continue to be responsible in our work on the important issues that AI travels and will find different ways to get out of these issues.

The panel should add external perspectives to ongoing AI ethics work from Google engineers, all of whom will continue. Hopefully, the cancellation of the board is not a withdrawal from Google's AI ethics work, but a chance to consider how to engage constructively outside of stakeholders.

The board became a big responsibility for Google

The credibility of the board first hit a hit when Alessandro Acquisti, a privacy researcher, announced on Twitter that he went down and claimed: "While I am devoted to research who intervenes in key ethical issues of justice, rights and inclusion in AI, I do not think this is the right forum for me to engage in this important work. "

At the same time, the petition to remove Kay Coles James has attracted more than 2300 Signatures from Google employees so far and showed no signs of losing steam.

As the board's mind intensified, board members were drawn into extended ethical debates on why they were on the board, something that Google could not have hoped for. On Facebook, board member Luciano Floridi, a philosopher of ethics in Oxford, the musket:

Asking [Kay Coles James’s] advice was a serious mistake and sending the wrong message about the nature and goals of the entire ATEAC project. From an ethical perspective, Google has misrepresented what it means to have representative views in a broader context. If Mrs Coles James does not resign, as I hope she does, and if Google does not remove her (https: //medium.com/…/ googlers-against-transphobia-and-hate -…), as I personally have recommended, the question becomes: What is the right moral attitude to take in light of this serious error?

He ended up joining the panel, but it was not the kind of ethical debate Google had hoped to spark – and it became difficult to imagine the two working together.

That wasn't the only problem. One day ago I argued that the noise was not well set up to succeed. AI ethics boards like Google's vogue in Silicon Valley do not seem to be well equipped to solve, or even progress, difficult issues of ethical AI progress.

A role on Google's AI government was an unpaid, toothless position that may not, in four meetings over a year, come to a clear understanding of everything Google does, say, provide nuanced guidance about it. There are urgent ethical issues about the AI ​​work that Google does – and no real avenue where the board could address them satisfactorily. From the start it was poorly designed for the target.

Now it has been canceled.

Google still needs to find out AI ethics – just not like that

Many of Google's AI researchers are active in making AI fairer and more transparent and clumsy missteps by management will not change it. The Google spokesperson I spoke to pointing to several documents that apparently reflect Google's approach to AI ethics, from a detailed mission statement describing research, they will not start looking back at the beginning of the year if their AI jobs so far produce socially good detail articles on the state of AI control.

Ideally, an exterior panel will complement that work, increase accountability, and help every Google AI project be subject to appropriate scrutiny. Even before the rebellion, the board was not set up to do so.

Google's next External Responsibility Staff must address these issues. A better board can meet more often and have more stakeholders engaged. It would also make public and transparent recommendations, and Google would tell us if they had followed them and why.

It's important that Google get this right. AI capabilities continue to advance, so most Americans are nervous about everything from automation to data security to catastrophic accidents with advanced AI systems. Ethics and government cannot be a side-show for companies like Google, and they will be under intense control as they try to navigate the challenges they create.

Sign up for the future Perfect newsletter. Twice a week you get an overview of ideas and solutions to tackle our biggest challenges: improving public health, reducing human and animal suffering, easing catastrophic dangers, and – to put it simply – getting better at doing good