The company created by PayPal co-founder Peter Thiel opens on the stock exchange between accusations of opacity and operations outside the legality and privacy of citizens
What happens if you mix an unscrupulous megamillionaire, able to secretly fund the bankruptcy of a media outlet, with artificial intelligence technology and contracts with some of the world’s major intelligence agencies, defense departments, and governments, from the CIA to the Pentagon to the FBI or the Centers for Disease Control and Prevention (CDC)? The result is Palantir, one of the most opaque and power-wielding technology companies in the world that is about to be released on the Stock Exchange today. Your departure to parquet on Wall Street will mean a before and after to see how far you can go in using algorithms for all kinds of tasks: from hunting down a terrorist to accepting or denying a mortgage in a discriminatory way.
Palantir’s IPO is one of the techno-financial events of the year. The company will do so unusually, without issuing new shares as happens in a traditional OPV. Instead, it will go straight to the parquet, as has already happened with other technology such as Spotify or Slack. And that’s where the party will start… or the funeral. Several analysts put the value of the company (which has never made a profit) at $22 billion, but others put it in less than half. But this is the least of it. The figures are just a reflection of the real ‘quid’ of the matter: does Palantir have an artificial intelligence technology worth so many millions?
The company, created by PayPal co-founder Peter Thiel, one of the first Facebook investors and Donald Trump’s colleague, says yes. His name is a statement of intent. The ‘palantir’ is the crystal ball of ‘The Lord of the Rings’ that allows you to connect with distant people and moments in time and space and spy on everything. Sauron uses it to see the past, keep an eye on his enemies and drive his victims crazy. Your product is easy to understand: AI ‘software’ capable of delivering real-time data that easily helps you make better decisions on the go.
Where are the most dangerous conflict zones that soldiers should avoid? Which voters do you have to touch to win an election? Which areas have the highest rates of covid-19 contagion? They are all lands on which Palantir has been helping the US government especially for the past few years. The problem pointed out by engineers and artificial intelligence specialists is that their technology is not necessarily better, it can fail and, above all, it is used with such opacity and secrecy that it is impossible to ensure that there are no serious illegalities in the form of discrimination or serious privacy breaches.
“I came out of there convinced it was easier to use.” This was recently explained to New York Magazine by Heidi Shyu, former head of arms purchases for the U.S. Army. Shyu tells how in a meeting with senior Palantir executives these showed how a couple of clicks were enough to create heat maps that point to the most dangerous combat zones. It is also something that made the existing ‘software’ of the army, but it took much longer. And Palantir’s didn’t hang as often as his competitors’ either.
This is how since 2003, aggressively and offering free trials to soldiers, Palantir began to be used within the U.S. military. It became so essential that the Pentagon had to take letters in the matter for negligence in the procurement process and excluded Palantir from the award for the purchase of this system. Thiel went crazy: he sued the Army and won the judicial pulse in October 2016. But that wasn’t the best thing. Just a few days later, his friend Donal Trump became president of the United States. Coincidentally, four months later, Palantir won what he longed for: a 10-year contract worth more than $876 million to develop the American army’s combat analysis software. The US Army already had one on which more than $2.3 billion had been spent. Money thrown away to replace it with Palantir.
It is just one of countless contracts the company has signed with companies and government bodies to place its two flagship products: Palantir Gotham and Palantir Metropolis. They’re named after the cities of Batman and Superman. Theoretically, Gotham is designed to monitor terrorist targets and analyzes relationships between people, places and objects, while Metropolis observes behavior over time to investigate illegal transactions linked to financial fraud but also to human trafficking, substances, stolen art, etc. In practice, as El Confidencial published last year in an extensive report, the two are tools to monitor people and are designed for three very specific industries: government, financial sector and legal sector. Both are predictive technology. Your job is to use data on what has happened to point out crimes that have not yet happened.
In the Palantir government sector it has been more successful, but Palantir Metropolis has been a fiasco in the financial sector. Alex Karp himself, co-founder and CEO of the company, stated internally that Metropolis was an “absolute failure”. Despite having Donald Trump’s favor and spending millions in lobbying in the highest political spheres, the reality is that Palantir continues to lose money in shovels (580 million last year). And the fundamental problem is its own technology: doing everything Palantir promises and doing it effectively and, above all, legally, is virtually impossible.
Like Cambridge Analytica (CA), Palantir uses public data, purchased data and third-party data. Unlike CA, it has access to the archives of federal, state, and local governments and their institutions. That is, their databases are full of arrests, complaints, police reports, fines, license plates registered by road radars… A 2015 ‘leak’ speaks of up to 160 different databases from the LAPD alone, including hot zones in difficult neighborhoods.
Numerous studies since its implementation in cities confirm its toxicity and the danger of its sesgade algorithms. Chicago’s “strategic list of suspects” has actually proven to generate more violence than it prevents. To top it all off, Palantir has access to sensitive personal data, such as fingerprints and other biometric data, medical records, card purchase history, travel logs, communications, taxes, child history. And it gets all the data it processes, for use with other customers. He takes his algorithmic bias wherever he goes. Thanks to several leaks we know that they have contracts with intelligence agencies in England, Australia, New Zealand and Canada.
Who controls all this? One man: Peter Thiel. As Bloomberg reveals, a change in the company’s governance model gives Thiel greater power in voice and vote than any other investor in the company, and gives it to him perpetually. Not a rarity in Silicon Valley, other companies like Facebook and WeWork have used a similar system to give their founders full power. And the potholes that Facebook and WeWork have dared (and traverses) are well known. The big unknown now is to see if Palantir operates under strict law enforcement or we are facing a new Cambridge Analytica-style scandal. The IBRO may be the ultimate needle that the smoke ball even though, unfortunately, until a new Edward Snowden or a Christopher Wylie dares to take the step, we probably never know.