Why each of us need to act now on Artificial Intelligence!

Context

The goal of this article is to make everyone conscious of the recent advances in Artificial Intelligent and its implications. I am citing some names and statements only to corroborate that reality. The companies and leaders in the AI domain are making important contribution to advance the technology, but not everyone’s goals align with what’s best for the humanity.

The race

There are systems that are being trained to read the entire humanities’ work in any language. Going even further, multi-modal models are being created that will go beyond language and ingest audio, video, images etc. The plan is then to feed these models with everything humans have generated.

In April 2023, Elon Musk demanded that the Artificial Intelligence research be paused. Last week he ignored his own call for a slowdown and open sourced the largest LLM at 314 billion parameters. In a recent interview Sam Altman, CEO OpenAI, said, Artificial General Intelligence (AGI) could arrive within the next five years. Jensen Huang, founder and CEO of NVIDIA, said that within the next ten years, the data centers will become AI factories producing things that did not exist before.

Goal misalignment

While every AI company paints the picture of solving complex problems to serve the humanity, they all need to monetize it immediately to fund the exorbitant cost of running and advancing the technology. Here are examples of common use cases that can be found on their websites, marketing material, and pitches:

• Replace repetitive drudgery being performed by highly paid information workers.
• Write prose and poetry; create images, drawings, paintings, etc.
• Automate repetitive tasks such as data entry and factory work.
• Mimic our activities and automate a lot of what we do.
• Watch humans, mimic them, and perform the task at the same level of efficiency.
• Automate time-consuming and monotonous workflows.

First misalignment is the goal to replace humans by mimicking what they do and attempt to do it better and faster. Yes, we have been automating with computers for the last 78 years since the world got its first general purpose computer in 1946. But the magnitude, scale, and acceleration that is coming ahead of us is monumental. We will experience unprecedented change in one tenth of the previous time span.

In his GTC Keynote speech, Huang used the word “accelerate” sixteen times in a span of five minutes. Huang also went on to say that the goal in life of these AI factories is to generate revenue. Second misalignment is the goal to maximize financial metrics.

Several companies have launched products that generate amazing graphics and mind-blowing videos. One of the speakers also talked about a full feature length show or movie generated entirely with AI. When we read books, watch movies, listen to music, or see paintings we make an emotional connection with the artists, writers, directors, actors, etc. Third misalignment is the goal to eliminate that emotional connection.

What purpose do these goals serve in the advancement of humanity?

Current Reality

According to a World Bank report, around the world, almost 700 million people live in extreme poverty today, earning less than $2.15 per day. OpenAI’s own research says that “19% of workers may see at least 50% of their tasks impacted.” Being a sensitive subject, very few corporations are going to openly admit that they are laying off because of AI. Recently, IBM announced that 8000 jobs will be replaced with AI. Google announced that hundreds will be laid off from its ad sales team; the team is expected to rely heavily on machine learning to automate advertising . Duolingo turned to GPT-4 and cut 10% of its contractor workforce. These announcements are going to have cascading effects. Further, this will disproportionately impact emerging and undeveloped economies.

Unrealistic Expectations

According to a World Economic Forum’s (WEF) The Future of Jobs report “44% of workers’ skills will be disrupted in the next five years.” It goes on to say that “six in 10 workers will require training before 2027, but only half of workers are seen to have access to adequate training opportunities today.” Beyond workplace, our societies and education systems are not equipped to produce the workforce needed for these new jobs in the next 3-5 years.

Majority of humans easily fall prey to spam emails, frauds, fake news, etc. If these AI-based systems are unleashed at an accelerated pace, how do we equip the global population with intellectual capabilities to make unbiased, fair, ethical decisions when using such systems?

Existential Threats

• Economic threat. These job losses and skill gaps will accelerate economic inequalities between the rich, middle class, and the poor. It will significantly widen the gap between countries with a large pool of technical talent versus those with limited talent.
• Threat to liberty, equality, and justice. The acceleration and scale could lead to further concentration of wealth with a handful of corporations. These corporations could use the wealth to lobby governments for favorable policies. To force these policies upon its citizens, we could see rise in authoritarian governments.
• Threat to privacy. Currently our personal data, confidential data, and intellectual properties are off limits for AI models. Bryan Catanzaro, NVIDIA's VP of Applied Deep Learning Research, stated that "the most valuable tokens" are the most closely guarded data.
• Threat to human diversity. We have been experiencing cultural extinction for centuries, but the AI race will only accelerate that process of wiping out languages, traditions, and indigenous tribes.

Call for Action

The future of humanity depends on our collective action. Only 30,000 individuals, AI experts and executives signed the letter to pause AI. It didn’t have any material impact. I am not advocating for a pause, but for goal alignment. The argument is not whether these advances will happen in ten years or twenty. It is also not whether it will impact 20% jobs or 10%. The rate of change will be much faster than any of us are prepared for. Several folks are celebrating the short-terms gains and ignoring the long-term consequences. Each one of us must step up and act.

Take the pledge to only enable AI or use it for self-development and to aid humanity. Say NO to everything else.

Disclaimer: The views expressed in the article are my own and do not represent that of Kyndryl.