Google just changed the rules on AI ethics – guess what’s back in play?

How much is a principle worth? In big tech, morality isn’t a code – it’s a price tag.

Google's decision to retreat from not using AI for surveillance and weapons manufacturing doesn’t come as a big surprise in the current political climate. It’s not the first time a tech giant has shifted its moral compass to align with the times.

Financial pressure and political influence play crucial roles in understanding the shifting sands of morality, but from a user perspective, when companies change their minds faster than a software update, it becomes difficult to predict what’s next.

The term “flexible ethics” sounds appealing, but in practice, it often means rewriting guidelines when it benefits the company. Google is no stranger to this, and neither are some of the other big players.

In 2018, Google used AI to analyze drone surveillance for the US military in an operation named Project Maven. After employees protested, arguing it was a step toward autonomous weapons, Google dropped the contract. But this week’s backtracking on its AI policy raises the question: will similar military partnerships return?

Another example of ethical drift occurred when Microsoft's cloud AI contract for the Pentagon was canceled in 2021 due to concerns it could be used for warfighting. Yet, Microsoft continued pursuing military contracts in subsequent years – just under different names.

Facial recognition has also sparked controversy, with both Amazon and IBM shifting positions over time. In 2018, Amazon began selling its Rekognition facial recognition tech to US law enforcement but paused sales after MIT research showed the system disproportionately misidentified Black and Asian individuals, particularly women.

Following the George Floyd murder in 2020, the backlash from activist groups intensified. Amazon paused sales again but never fully withdrew the product. Instead, they quietly resumed service in 2023 with police and government agencies.

IBM, on the other hand, publicly distanced itself from facial recognition, announcing it would stop selling the technology. Yet, despite this positive PR move, IBM continued partnering with police departments like the NYPD and South Fulton Police Department in Georgia to expand AI-driven law enforcement tools.

Google is pouring $75 billion into AI this year, a 29% increase over expert projections. This surge in Gemini integration places Google among the AI leaders. While users gain a convenient assistant in Google Docs, is the company also accelerating its entry into AI warfare behind the scenes?

Google will frame its shift toward AI arms development as a values-based decision. However, as the US competes with China in AI warfare and surveillance, one paradox remains: National security is the justification, but in practice, it can lead to mass surveillance, biased policing, or autonomous warfare.

Google removed its "don't be evil” motto from its official code of conduct in 2018 (originally a dig at competitors). Even after changing it to “do the right thing,” the question remains – for whom?