Meredith Whittaker, President of Signal, spoke to Euronews Next about Europe’s regulation of AI, privacy and mental health in this new era of artificial intelligence (AI).
Meredith Whitaker, president of the nonprofit Signal Foundation, which operates the encrypted Signal messenger app, is one of the early thinkers on artificial intelligence (AI), having led a research group on the technology for Google in 2006.
He led a walkout over Google’s attitude towards sexual harassment and was ousted from the company in 2019.
She then co-founded the AI Now Institute and was an advisor to the US Federal Trade Commission and still has concerns about Big Tech and how it controls our data in this new age of AI.
He spoke to Euronews Next about AI and regulation in Europe at this year’s Web Summit in Lisbon.
EuronewsNext: You’ve spoken a lot about your fears about AI and all the AI hype we’re living through. What is it specifically that you are concerned about?
Meredith Whittaker: I think my fear is less about the technology and more about the fact that it is developed and controlled by a handful of large corporations, whose interests, of course, are the corporation’s interests and profit and growth. And to make people happy. Shareholders; Not necessarily social good.
And the AI that they’re selling, that they’re licensing, that they’re building, that they’re deploying, it requires massive amounts of data, requires massive amounts of computational power, And effectively reinforces and extends the surveillance business model, which is why much of the world is concerned about the pitfalls of so many when it comes to irresponsible technology.
I think my fears and my concerns are more about the people, institutions, and incentives that are shaping AI, rather than they are about the technology itself, or the idea that it is somehow sentient or Can become like God.
Let’s talk about governance and regulation of AI. Does regulation go far enough? And more specifically, as Europe prepares its own AI mandate, how is Europe faring?
[Europe’s] The AI Act is not complete yet. We don’t know exactly what the final form will be. I think we need to suspend judgment for a minute. I think Europe has already done a lot that can be used to powerfully prevent some of the harms of AI.
One example I would put forward is the fact that the GDPR is strictly interpreted and enforced. Strictness could easily be used to ban surveillance advertising, and banning surveillance advertising would have a major and quite radical restructuring effect on the incentives that drive these corporations.
This would mean that suddenly the huge amounts of data used to create demographic profiles that are sold to advertisers for micro-targeting, targeting different types of people, would no longer be acceptable.
And that will definitely impact AI because that data source, you know, everything we post on social media, our location data, our social graph, all of that will suddenly become much less useful.
And in a strict interpretation, perhaps archive and maintenance is not allowed for these centralized corporations. So, this is an example where Europe is already on this. The question is actually of political will.
Let’s dive into the political will, do you think maybe there’s this idea in America of let’s be more corporate about this, let’s make more profits?
Of course. A large number of these corporations are based in the US. This is their jurisdiction and America is well aware of this. These are corporations that partner with the US government.
Often they provide infrastructure for the government. And we’re in a very chaotic political situation in America, where it’s been 20 years, 30 years, depending on where you are, and we’ve had no federal privacy laws.
One thing that gives me hope is that it is very difficult to build these massive global-sized systems in a specific way for each jurisdiction.
Such powerful regulation from Europe or another major market has an impact on these corporations, even if it does not come from the jurisdiction in which they are located.
How dangerous will AI “sweatshops” that collect data?
I think the issue there is really a labor question. It was recently estimated that there are 100 million people employed in highly precarious positions who are effectively human managers, providing the human intelligence that AI and AI companies take credit for.
So that’s data labeling. That’s data curation, that’s calibration. Telling the system what is acceptable and what is unacceptable. That is content moderation; It is the error control and management of the problems that these AI systems generate when they behave erratically.
This is a very large number of people.
And I think there are real issues with the quality of that work. There are real issues related to the damage that work does to people’s mental health. This is not controversial.
The act of repeatedly viewing or encountering frightening images or texts has a documented, and very clear, detrimental effect on people’s mental health. We have testimony from people who have lost their jobs, lost relationships, and lost their ability to live in the world because of how disturbing it was.
And I think there’s a bigger question that we need to step back from, if these systems require losses on the scale that they do, to make these systems work and make them acceptable for business and popular use. There is no other way to make it.
Are we comfortable with this?
Do you believe that privacy will be one of the more significant disadvantages due to the increasing number of people using generic AI software like Chat GPT?
Privacy has long been a casualty of the technology business model.
Signal will continue to provide a truly private safe haven from the abundant data collection that defines the rest of the industry.
But of course, this race to be bigger is better. AI poses a huge privacy risk. It demands the creation and collection of increasingly intimate, increasingly pervasive, increasingly invasive data about us, our communities.
And it’s a power to further centralize that data in the hands of a handful of companies, which is again an irresponsible power. It is very difficult to find its analogue in history.
You have criticized the UK’s Online Security Act regarding the abolition of end-to-end encryption. How do you feel about this and the EU’s data rules?
[The UK’s Online Safety Act] This included what I would actually consider a product or at least partially a product of this AI hype, which was a very misleading section based on the belief that AI systems would way can scan everyone’s personal communications.
Matching it with a database of permissible expression in a way that is private and secure and, of course, a long-standing technical consensus. The entire expert community knows this is not true.
And so we fought against it because as Signal, we rely on end-to-end encryption to provide a meaningful option for truly private communications in a world full of surveillance. Because we were very concerned about that law.
Now, similar legislation is moving forward in Europe. It seems like they’ve adapted it to exempt Internet-encrypted messaging from scanning, which is great. But of course, trials are ongoing and there is a complex political process. We haven’t finished the work yet.
But we’re glad that by bringing forward the human rights community, bringing forward real technical expertise, not just the marketing departments of AI companies, we were able to make clear that this was magical thinking and actually exceptionally important for rights and freedoms. It was dangerous. ,