Humans’ value beyond AI

Imagine a world where AI systems write articles, answer customer queries, and even perform complex tasks once exclusive to human expertise. In this future, what will set us apart, and how can we ensure our continued relevance and success?

What value requires being human?

ChatGPT wrote the above opening statement and title of this article, and it’s on target. Now, why do you want to listen to me? I’m a Senior Machine Learning Engineer at a company helping employees learn skills, an entrepreneur with a Master’s in Computational Linguistics, and a certified hypnotherapist. In other words, I’m interested in both computer minds and human minds. I use ChatGPT and OpenAI APIs every day, and I’m here to tell you that you can best develop yourself professionally in a world of AI by understanding both computer and human minds.

I also have been a hiring manager at three organizations, and at the end of this article I’ll discuss how to evaluate the qualities I’m suggesting developing using strategies I’ve found effective.

What skills do humans have that AI doesn’t?

Let’s start by considering ChatGPT’s suggestions for the most important skills for humans to develop (though I gave it #4):

1. Mindfulness: Cultivating self-awareness, resilience, and empathy allows us to better adapt to new technologies, environments, and situations.
2. Emotional Intelligence: As AI takes on more tasks, the ability to manage emotions, understand others, and communicate effectively becomes increasingly important.
3. Creativity: AI may excel at generating output, but humans possess the unique ability to think outside the box, innovate, and solve complex problems.
4. Outcome Specification: Learning to define outcomes in detail and with clarity is crucial, as it enables us to communicate our goals effectively and harness the power of AI.

I’m actually going to challenge this answer a bit on points 2 and 3. AI can and does have both emotional intelligence and creativity to some degree, but humans have access to a sort of intuition that AI probably won’t for a long time.

Let’s consider two theories on human consciousness:

1. A common theory in research of psychic abilities like ESP is that humans have a “spiritual” consciousness that’s independent of the brain (see one of these papers on a non-local universe or non-local consciousness), receives information from other consciousness, and then informs creativity and emotional intelligence.
2. A more mainstream theory is that consciousness arises from the brain. In this case, there is nothing inherently stopping AI from eventually being as capable as the brain; therefore, in terms of intelligence, we would have by definition nothing unique to add other than our personal experience with our body which could be simulated by an AI, since they’re trained on humans with bodies. (This theory would also mean that consciousness can be created mechanically.)

Creativity could be described as coming up with something new that’s useful. As an example of creativity from ChatGPT, I asked it to come up with a skill for which there is no term, and it suggested the below:

“contextual emotion adaptation” in the field of customer support. This potential skill involves intuitively adapting one’s communication style and tone based on the emotional state of the customer, using cues from their language, expressions, or behavior. While professionals in customer support may already be practicing this skill to some extent, it hasn’t been explicitly recognized or named.

“Contextual emotion adaptation” has 0 results on a Google search.

The key point here is that AI has few limits. The recent article “GPTs are GPTs: An Early Look at the Labor Market Impact” by OpenAI, OpenResearch, and the University of Pennsylvania says that 19% of workers may see at least 50% of their tasks impacted by AI, and 80% may see 10% of their tasks impacted. So what do we do about that?

Inner Experience vs Output

Let’s assume that we are conscious and AI is not. This being the case, we have a unique ability to explore and enhance our consciousness and inner personal experience. This is also known as “developing our consciousness”.

In education and especially public schools, students are evaluated by test scores. This is output. Now AI is already better than most humans at taking tests (see https://openai.com/research/gpt-4) and could match humans in all areas as early as next year (Artificial General Intelligence), so are tests really the best way to evaluate humans, or the most important outcome to measure? I don’t dive deep on education in this article, but my suggestion is for each student to have their own outcomes and for teachers to focus on helping each student define for themselves what outcomes they want for themselves, with happiness and motivation as the principal measures of success.

Output:

1. Language and communication
2. Code
3. Doing things

Inner experience:

1. What words can’t fully describe
2. Who outputs are intended to serve
3. Intent

Taking this a step further, even thoughts and mental visualization are examples of output, though they’re more closely tied to inner experience. First of all, we tend to think using language, and we can consciously bring up or remove thoughts and mental images (a useful skill to practice, by the way).

Meditation and mindfulness practices teach us how to focus on consciousness that is independent of thought, and see ourselves as distinct from our thoughts. This is a great step towards developing our consciousness through being meta-aware of our inner experience.

Conscious Focus on Outcomes

Recent advances in AI provide us a fantastic opportunity to spend more of our time focusing on outcomes, and less on the implementation details. If you want to rocket your way to success in a world of AI, learn how to do the following:

1. Want something that’s meaningful to yourself and others.
2. Experience the future outcome that you want (see and feel what you want, in vivid detail; hypnotherapy techniques are great for this).
3. Communicate the outcome you want to AI, leveraging an awareness of what assumptions and context you have that the AI doesn’t. This skill can be developed by communicating outcomes with humans as well, which is similarly useful.

The above could also be described as “manifestation + AI”. Ourselves and the world around us tends to move towards the outcomes that we focus on. ChatGPT or AI is simply a tool that allows us to achieve those outcomes more easily, freeing us up to spend more time deciding what we want those outcomes to be.

To be clear, this means telling ChatGPT something along the lines of “These are the standards for what I want you to do for me” instead of just “Do this for me”.

Example of a prompt I sent to ChatGPT:

“Desired outcome: I and others understand how humans can provide value distinct from AI such as ChatGPT and related APIs.

I publish a Medium article about this topic, including:

1. How to leverage AI in our job.
2. What the most important skills or tendencies are for humans to learn for their career now that AI has advanced.
3. How we, education, and society tend to focus too much on output, including words. AI is effective at output, but humans have conscious intent and experience, of which output is just one manifestation. If we focus on improving the inner experience, we can also improve other outward results as a side-benefit.”

What it means to add value

The most important value we can add is the addition of our own personal consciousness to whatever role, organization, or client we’re involved with.

What we put our mind to is more likely to happen, and AI, words, code, etc. are tools our consciousness can use to accomplish the outcomes we focus on.

Let consciousness, not time, be the new currency of work. Specialization and experience is still relevant, in this case, as it affects the quality and speed of our ideas and our ability to bring those ideas into reality. In other words, we’ve trained our brains to work effectively with our consciousness, and our consciousness to work effectively with our brain. The brain by itself is not so useful relative to AI; it’s how they work together that enables us to add value to the physical world.

Evaluating human consciousness

I will describe a few ways to measure the “quality of consciousness” that someone is bringing to your organization (and which you can apply to yourself).

Contribution statements

A contribution statement allows applicants who are particularly passionate about your company to stand out, and provides an opportunity for them to demonstrate critical thinking because they have to adapt their statement to the organization (it’s also a useful exercise for any new hire).

Evaluation criteria:

1. Unique to the person, such that it wouldn’t apply to any other person.
2. Unique to the organization, such that it wouldn’t apply to any other organization.
3. The above two points are effectively combined for every point in the statement.

Here’s an example short contribution statement written by ChatGPT, given the above criteria, a request for clear examples for both experience and contribution, a job title, and a company name:

“As a distinguished Senior Machine Learning Engineer with a wealth of experience in creating impactful solutions, I am excited to join Cornerstone OnDemand and apply my unique expertise in the following ways:

1. Utilizing my patented “AdaptLearn” recommendation system, I will create a tailored learning experience for Cornerstone’s platform users, ensuring content aligns with individual needs and preferences, thus boosting engagement and retention.
2. Implementing the award-winning “FairTalent” algorithm, which I developed to reduce hiring bias, I will work to enhance Cornerstone’s talent management processes, promoting diversity, equity, and inclusion throughout the organization.
3. Applying my cutting-edge NLP research, published in the “AI for Education” journal, I will develop innovative knowledge-sharing tools for Cornerstone’s collaborative learning environment, fostering seamless peer-to-peer interaction and more effective learning experiences.
4. Drawing on my successful mentorship of 15+ junior machine learning engineers, I will provide guidance and support to Cornerstone’s technical team, fostering growth and skill development while nurturing a collaborative and innovative culture.

I still think a contribution statement is a worthwhile exercise and insightful, but this statement is better than the statements I’ve read from most humans and it’s acceptable to me because of how specific and unique it is. Let’s move on to the next test.

Unassigned Innovations

It’s one thing to be able to complete a task, and quite another to come up with the right tasks to do given the context of the ongoing situation. Can ChatGPT do this?

Query:

Let’s say you’re responsible for Cornerstone’s learning content recommendation system but were not provided any tasks. What tasks would you do?

Response:

“As the Senior Machine Learning Engineer responsible for Cornerstone’s learning content recommendation system, I would initiate the following tasks to enhance the system and deliver a more personalized user experience:

1. Analyze existing data: Evaluate the current recommendation system’s performance, identify areas of improvement, and examine user interactions to better understand their preferences and learning behaviors.
2. Feature engineering: Identify and extract relevant features from the available user and content data, such as user demographics, past learning history, content metadata, and user engagement patterns, to improve the recommendation model.
3. …”

ChatGPT proceeded to provide 8 example tasks, and a clear explanation for how they achieve the platform’s overall objectives.

In the examples I’ve provided of ChatGPT’s output, it was able to reach success because of my guidance, but we need a human to make a specific request in the first place, right? Not necessarily, because it can be trained on human prompts and even automatically decide when to come up with something new to do. It can call other technology to carry out its ideas (see ChatGPT Plugins).

Review AI Prompts

Essentially any language can be replicated by a large language model, so one option to evaluate a human’s contribution is to request they send you their prompts to the AI, along with the responses. This way you can clearly distinguish what they contributed vs the AI and evaluate their ability to come up with ideas, specify outcomes, and leverage AI.

Again, even both sides of the conversation can be created by ChatGPT, so it’s not distinctly human. That being said, from personal experience, I’ve found that humans that don’t put much consciousness or effort into their request to the AI often don’t have high quality output.

Here’s a short example of this from ChatGPT, where the user’s prompt is not generic and sounds different from what I would expect from its training data:

User: Design a creative fundraising event for a small animal rescue organization that incorporates elements of technology and social media.

ChatGPT: How about organizing a “Virtual Pet Talent Show” to raise funds for your small animal rescue organization? This unique event can engage your online community while promoting your cause. Here’s a basic outline:

Do it live?

Of course, you could always evaluate humans live, and I think this is useful, but I’m talking about hiring people to achieve outcomes on their own, and if we can’t distinguish between their work and an AI’s work, what are we hiring them for?

By improving our hiring criteria, educational objectives, and what we value in human ability and the workplace, humans will adapt to focus on how they can uniquely provide value in a world of AI. This shift in human development will result in a better world due to more value creation and development of our own inner consciousness and awareness, which in turn would result in more unconditional love.

Unconditional Love

The concept of “psychological conditioning” is directly related to training an AI. An AI can be psychologically conditioned similarly to us, so if we want to distinguish between what we consciously bring to the table that’s different from AI, that means being aware of and rising above our psychological conditioning. This is essentially what meditation is, being aware of what our brain is doing instead of being.

Even our emotions are conditioned and greatly affected by our brain. For this reason, I don’t doubt that AI can replicate our emotions as well (though experiencing them is something else).

Because we are all psychologically conditioned, most likely everyone’s love is conditioned in some way. No wonder, then, that those who practice meditation (quieting our thoughts) often report experiencing unconditional love, as if the conditioning of our brain is getting in the way.

Since an AI is only intelligent as a result of its training, also known as conditioning, as humans we can add value through unconditional love. Developing our unconditional love puts us in a position to best guide and use AI, and of course make the world a better place.

Is the human brain similar to AI?

Next Steps

In summary, to be competitive in a world of AI employees, you can develop the following capabilities:

1. Be aware of your conditioning and condition yourself more towards unconditional love (including for yourself).
2. Take control of and train your own subconscious and psychological conditioning (e.g. through self-hypnosis or hypnotherapy).
3. Bring consciousness to your work, which results in thinking outside of the box.
4. Receive and act on insights (consider learning about the Silva method, the science of channeling, or non-local consciousness in general).
5. Practice visualizing and communicating outcomes.
6. Come up with your own outcomes and ways to provide value (this is aided by the above points).

Thanks for your attention.

PS: There was a lot to cover in this article and I’d love to dive deeper. Please let me know in the comments what you’d like me to elaborate on for future articles!

PPS: GPT-4 was used for all ChatGPT outputs, and the images were AI-generated by Fotor.