Introduction – Prepare against Perils of Vicious Artificial Intelligence
We hear “AI has both a bright side and a dark side. We should enhance the former and weaken the latter for developing responsible AI”
We cannot disagree but would like to add one more remark to it – “We should be prepared to fend off bad guys who attempt to exploit the dark side of AI”.
We all know that we cannot eliminate bad guys. We only can build the defense line against them. We must prepare against vicious AI deployed by malicious guys for malign objectives. Among the malicious guys are dictators and autocrats, and among the malign objectives is destruction of the values of democracy.
The domain we are active in is Digital Identity. Naturally we have an obligation to consider what we can do for the defence of democracy from digital identity’s view point.
AI Program with No Identity
AI might become extremely excellent in mimicking human thinking. Mimicked thinking, however, is just different to human thinking.
Human thinking is supported, reflected, influenced and inspired by humans’ internal/external body senses and emotion, feeling and mood that come from them, as well as history of social and personal experiences and time-honoured wisdom born from all of them. ‘Identity’ is made up of these elements which can hardly be an object of mathematical calculations.
Mimicked thinking may be feasible without Identity, but human thinking is never possible without Identity. Each separate AI program, if given an ‘identification code’, can never be an object of ‘Identity Authentication’. Should we mess up here, we might be walking on the road towards a digital dystopia.
We have to make it clear that we are facing the AI that does not have and will never obtain Identity.
Marriage of Artificial Intelligence and Doublethink
What would we see when someone combines a truly powerful mimicked intelligence with Doublethink? – It would be the worst possible form of digital dystopia.
In a world where AI-integrated Doublethink dominates, it should be viewed as correct if someone states that a surface contained in a larger surface can be larger than the larger surface that contains it.
It enables them to assert that an attack surface is larger than a defense surface in the case of the password while a defense surface is larger than an attack surface in the cases of all the other security factors.
We know that what is technically feasible have happened. There are always people who make it happen. That is, we will have to see a happy marriage of Doublethink and Mimicked Intelligence at any time soon.
What if Doublethink gets blown into Identity-less Humanoid?
“Professionals across the tech landscape have expressed concerns over Microsoft’s plans to extend the capabilities of ChatGPT to robotics and controlled multiple platforms such as robot arms, drones, and home assistant robots intuitively with language.”
This is the message found in this report by Market Edge Online – https://marketingedge.com.ng/pay-critical-attention-to-developments-in-robot-technology-experts-warn/
It may not be long before someone starts to blow ‘Doublethink’ into Identity-less Humanoids. It is thus getting visibly obvious that the issues of digital identity is inseparable from the issues of artificial intelligence.
Few people are born to be receptive of Doublethink blown into their brain. It would require intimidation or brute violence to force the people to accept the doublethink in most cases. Does the same apply to Identity-less Humanoids?
Artificial Intelligence as Nemesis of Biometrics
The phrase ‘Nemesis of Biometrics’ occurred to me when I came across this Washington Post report – “They thought loved ones were calling for help. It was an AI scam.”
https://www.washingtonpost.com/technology/2023/03/05/ai-voice-scam/
It reads “Scammers are using artificial intelligence to sound more like family members in distress. People are falling for it and losing thousands of dollars.”
With biometrics conveniently waiting there to be abused, it’s only natural for criminals to look to AI as a nice tool for easier and quicker impersonation and identity fraud. This topic of ‘artificial intelligence and digital identity’ is really intriguing indeed.
I would also like to take up this article “Here’s How We Broke Into a Bank Account Using an AI-Voice” https://www.vice.com/en/article/qjkgwx/cyber-podcast-how-we-broke-into-bank-account-ai-voice
This event demonstrates that False Acceptance Rate of the bank’s voice verification was not set as low as desired.
The solution? -Lower the FAR down close to the absolute ZERO and we would not see this type of dreadful events. We would , however, be witnessing a very frequent case of False Rejection.
A typical case of AI as a friend of bad guys? Or, a typical case of biometrics in the AI-era?
Digital Identity Racing with Artificial Intelligence
What would happen to the password when AI becomes more powerful?
Firstly, we do not need to wait for the arrival of truly powerful AI for the poor fate of passwords. Once leaked, any strong passwords stored in plain text would not need any computing power. Even when hash-protected, passwords of low entropy can be broken by conventional computers.
Secondly, when AI becomes truly powerful to break very strong passwords, it would really be a grave threat indeed if we have no measures to fend it off.
We would not have to despair, though. We still opt to keep making the entropy of our passwords ever higher.
It’s obvious that our brain can by no means handle it and, however powerful, passwords written/stored on memos/devices are just vulnerable to theft. But we have a solution: get extremely-high-entropy passwords generated from image data as illustrated below.
The entropy of several images combined can easily exceed a million bits. What we need to do, on the other hand, is only locating several memorable images associated with our unforgettable pleasant experiences.
Memories of our personal experiences are outside the realm of artificial intelligence running on someone else’s computers.
Remark: Threats of ‘visual-manual attacks on display’ are very different to ‘automated brute force attacks’ on the data server.
A figure of ’20-bits’, say, a million attempts, for instance, would be just a bad joke against automated attacks, whereas it would make a pretty tall wall against visual-manual attacks on display.
Modest AI Developers Wanted
As mentioned earlier, human intelligence is supported, reflected, influenced and inspired by people’s history of personal experiences, internal/external body senses and the emotion, feeling and mood that come from them, while artificial intelligence does not have and can by no means obtain any of them.
What this tells is that artificial intelligence, however good at mimicking human thoughts and behaviours, is and will remain Mimicked/Pseudo/Quasi Intelligence forever. So long as AI developers claim that their brainchild will replace human intelligence, what we see would be no more than the extremely costly disturbances, hypes and chaos.
However, once AI developers acknowledge that artificial intelligence can never match human intelligence in the real sense of intelligence, it could be a friendly force that might help us in the real way. We expect such modest AI developers to help us for our endeavour of building a sustainable digital identity platform.
Defense Line against Privacy Risk Resides in Citizen’s Brain Unnoticed
While we are unable to keep AI-armed bad guys away off our privacy altogether, we should be able to deter those bad guys from accessing our digital privacy data.
Key is what is inscribed deep in citizens’ brains as episodic memory that is unique to each person and stay outside the realm of any powerful artificial intelligence running on someone else’s computers.
What solves the password headache has long existed in citizens’ brains as autobiographic memory, especially episodic image memory. People simply were not awake to its power. The power of citizens’ non-volatile memory of their personal experience also makes the defense line against the AI-armed adversaries of privacy
Conclusion
What makes humans’ identity can make a defense against the perils of Identity-less Artificial Intelligence. What can contribute to the solid digital identity platform can also contribute to the solid defense line against the perils of vicious artificial intelligence.
