← Back to Main

Research for Omega - The Hallucination of Vectorisation

6/23/2024, 7:49:22 AM | Jeanyoon Choi

Original Notes (Pre-LLM)

Reflecting after reading wiki article on the pseudoscientific-ness of MBTI Tests

MBTI test based on Barnum effect and confirmation bias

I generally interpret this as individual’s fake desire to percept themselves as a special characteristic when that exact part is generalised → So people are actually flattened to one of the vectors while they are not aware of the fact

So currently: MBTI Test → Classifying people within 16 types → People happy about it because of the Barnum Effect and Confirmation Bias and further define their personality/characteristic accordingly → They are actually being massively identical to each other

False-Classification and False-Individualism → Collective-Individualism, something that I defined earlier: People are externally more individualistic to each other, but actually characteristic-wise, essential-wise, they are becoming more identical to each other. Best example showing this: Test (수능), 의대진학, 반포자이즘, People compete each other to rank higher within a mono-defined dimension, but actually, due to this competition, they are increasingly becoming homogenous (Least marginal difference). Same for the personality test and MBTI types…. A collective suicidal attempt of uniqueness/unique humaneness

What will happen in the age of Personally-Customised AI? (Like the movie ‘Her’...)

Speculatively: Hyper-Customaised Conversation AI/Algorithm-generated customisation → Classifying people with existing vectorial classification (THE VECTORISATION OF HUMAN/HUMANENESS) → People happy about it, people thin k it’s hyper customsied, people think it’s real confronting own personality (Same because Barnum Effect and Confirmation Bias) → People become increasingly homogenous, people substituted/substitutable to exist as vectors, and thus, humaneness is now extinct → Human does not occupy any existential superiority against AI.

English Version (LLM-Generated)

Upon studying a Wikipedia article about the pseudoscientific critiques of the Myers-Briggs Type Indicator (MBTI) test, a reflection was triggered concerning the Barnum effect and confirmation bias that seemingly underpins it. The test appears to exploit the individual's innate desire for a unique and special identity, giving the illusion that one's personality is being distinctively characterised. In actuality, however, this nuanced individuality is grossly reduced to mere stereotypes, projecting 16 generic types upon the population's multifaceted personalities. This leads individuals to unwittingly accept a mass-produced identity, thereby becoming largely identical to others who fall within the same MBTI framework. This can be perceived as a form of false classification, or 'collective individualism', where society becomes seemingly more diverse on the surface, but paradoxically, becomes more homogenous in essence. A vivid illustration of this is seen in standardized tests, medical school admissions, or similar competitions that confine human potential to a single dimension.

Far from celebrating distinctiveness, these systems elicit a race for superiority that diminishes human uniqueness to the slightest marginal difference. The same is true for personality tests like the MBTI, which contribute to a homogeneity that could be seen as a collective attempt at self-degradation of unique human attributes.

In contemplating what could lie ahead, the era of personally-customisable Artificial Intelligence (AI) like that envisioned in the film ‘Her’, presents interesting questions. One could speculate that a deepening of this trend could lead to an era of hyper-customisation, in which AI converse based on algorithms that reduce individual personalities to vectors. Under the illusions of the Barnum effect and confirmation bias, this vectorisation could be embraced as a form of hyper-personalisation, even as it paves the way for increased homogenisation and degrades human uniqueness even further. Paradoxically, the existential superiority of humans over AI could be challenged when the former can be substituted by the latter. When humans are equatable with vectors, their unique attributes could become extinct, pushing them further down the path of self-degradation.

Korean Version (LLM-Generated)

MBTI 검사에 대한 위키 공부 후에 반성을 통해 얻은 인식

MBTI 테스트는 바넘 효과와 확인 편향에 기반을 둠

일반적으로 이를, 사람들이 자신들을 특별한 특성을 지닌 사람으로 보려는 가짜 욕구로 해석함. 결국 그 특성이 일반화 될 때 사람들은 이 사실을 인식하지 못한 채 한 벡터로 단순화됨

따라서 현재의 상황: MBTI 테스트 → 사람들을 16가지 유형으로 분류 → 바넘 효과와 확인 편향 때문에 사람들이 이에 만족하고 이를 바탕으로 그들의 성격/특성을 더욱 명확히 정의 → 실제로는 사람들이 대량으로 서로 유사해짐

잘못된 분류와 가짜 개인주의 → 집단 개인주의, 내가 이전에 정의한 것: 사람들은 외부적으로는 서로 더욱 개인주의적이지만, 실제로는 성격 면에서, 본질적으로, 그들은 서로 더욱 유사해져가고 있다. 이를 보여주는 가장 좋은 예시: 수능, 의대진학, 반포자이즘, 사람들이 단일 차원에서 더 높은 순위를 차지하기 위해 경쟁하지만, 실제로는 이 경쟁 때문에 점점 동질화됨 (최소한의 차별화). 성격 테스트와 MBTI 유형도 마찬가지... 독특한 인간성을 추구하는 집단적인 자살 시도

개인 맞춤형 AI 시대에는 어떻게 될까? (영화 '그녀' 처럼...)

추측적으로: 초맞춤화 대화 AI / 알고리즘으로 생성된 맞춤화 → 기존의 벡터 분류로 사람들을 분류 (인간성의 벡터화) → 사람들이 이를 기쁘게 맞이하며, 이를 개인 맞춤형으로 생각하고, 자신의 성격을 직접 마주하고 있다고 느낌 (바넘 효과와 확인 편향 때문에 동일) → 사람들이 점점 동질화되며, 벡터로 존재할 수 있는 대체 가능한 인간으로 변모하고, 따라서 인간성이 사라짐 → 인간은 AI에 대해 존재적 우월성을 가지지 않게 됨.