Copyright: carballo

Electric cars spy on us

This month, the Mozilla Foundation published an ominous report with the title It’s Official: Cars Are the Worst Product Category We Have Ever Reviewed for Privacy. All 25 car brands reviewed collected more personal data than to operate your vehicle or manage their relationship with you. Most share or sell this data, and almost none gives you the possibility to delete the data. Nissan is the worst, stating that they can collect and share data about your sexual activity, health, and genetic makeup for marketing purposes. Read the analyses of privacy policies per brand here.

The report contains a glimmer of hope because two European car brands (both owned by Renault) did satisfy the European data protection regulation (GDPR). But they were vague about the data they did collect. This makes me wonder how Chinese cars, not investigated by Mozilla, but increasingly popular outside China, would fare under scrutiny. Cars are rich data collection machines.

AIs are networks

The report should remind us of the fact that modern, data-intensive AI is a network phenomenon. AI is not a single machine with an artificial brain. It is a network formed by our smartphones, electric cars and intelligent home devices, all of which provide machine learning algorithms with the data needed to identify patterns in our behavior that can be used to predict our behavior. This can be used to recommend us to buy stuff that people like us also bought.

In today’s intelligent networks, data is collected about us in return for some service; money is collected from us in return for some other offer.

Of course, instead of recommendations, the network could also send us warnings, fines, or reduced financial credit scores because its statistical model of us says so. It can exclude us from some consumer categories and offer us less attractive deals. It could give us a score that recommends additional surveillance, label us as a terrorist, or cancel our passports. This is seriously creepy.

Moderate, but lucrative intelligence

What is especially disconcerting is that Big Tech ¾the owners of the vast data collection, machine learning,  and people-nudging networks¾points our attention in the wrong direction by warning us of the risk of extinction by a superintelligent AI.

This helps to boost share value and attract smart engineers, but as a prediction of the level of intelligence it is plain bullshit: ChatGPT is at the end of its development. Its main function, producing grammatical, well-structured texts in natural language in response to a question, cannot be developed further, because its models cannot become larger. Current work on ChatGPT is aimed at preventing it from producing offensive texts and on defending its owner, OpenAI, against claims of intellectual property violation.

Instead of being afraid of the existential risk offered by superintelligent machines, we should be afraid of the existential risk posed by being treated as resources for extracting data first and money next. The Intelligence in this network does not need to be super; it just needs to make its owners more money. We find this moderate but lucrative intelligence in the networks of, among others, Alibaba [1], Amazon [2] and Google [3]. All you need is a gullible audience, vast amounts of data, and unbounded greed.

Proof of the existence of superintelligence: Pascal’s wager

The warning for superintelligence is similar to Pascal’s argument for the existence of God. His argument was:  I know that it is highly unlikely that God exists, but I cannot rule it out either. But if I don’t believe in the existence of God, then I would be seriously, seriously in trouble when He turns out to exist after all. It would mean that I enter hell after my death. To avoid the risk of eternal damnation, I’d better believe in God.

Big Tech says: It is highly unlikely that a superintelligence will exist, but we cannot rule it out either. But if we ignore the possibility and it turns out to come to pass anyway, then we are in deep, deep trouble. A superintelligence which aims at making paperclips without regard to other goals, will shut off all other machines and turn the whole world into a bunch of paperclips. Humanity will enter hell. To avoid this risk, we’d better believe a superintelligence will come to pass.

We need radical transparency of Big Tech

Let’s return to the real world. We are observed, classified, nudged, judged, admonished, warned, included and excluded based on data collected by our smartphones, cars, and smart home devices. If all we ever notice from this is that we see personalized ads and get our replenishments on time, then this is bearable, although it is still a huge waste of electricity.

But we notice a lot more. We are racially profiled, served conspiracy theories, submerged in hate speech,  and branded as terrorists. We are in deep trouble already.

Let’s turn around the searchlight 180 degrees and make the data collection and machine learning activities of Big Tech (including Big Cars) as transparent as possible so that they can be regulated in the interest of the public rather than that of Big Tech shareholders.