Jordan Peterson's Misunderstandings About AI

Recently, during an interview with Lord Conrad Black at the Canadian Democracy Fund, Peterson discussed ChatGPT, an AI chatbot developed by OpenAI. He mistakenly portrayed ChatGPT as having intelligence comparable to that of a human being, suggesting it would reach that level within a year or two. This is an exaggeration; ChatGPT is not designed for general intelligence. Its developers explicitly state that it is a specialized AI focused on generating text based on user input.

In simple terms, ChatGPT has been trained on a vast dataset of questions and answers, allowing it to generate responses based on statistical probabilities. It does not engage in actual thinking but simulates this process by predicting word sequences derived from extensive text training.

Peterson continues to misrepresent ChatGPT's capabilities, claiming it derives its analysis from human speech. This is incorrect; ChatGPT operates with text, not speech, and it is neither speech recognition nor synthesis software. Furthermore, he asserts that ChatGPT has not been trained on real-world data, which is also false. In fact, it has been trained on a diverse array of real-world text, including web pages and books, but its knowledge is limited to information available up until 2021.

As someone who has studied artificial intelligence as part of my computer science education, I am familiar with these concepts far better than Peterson. Experts in various fields have also criticized his inaccuracies, including biologists who challenge his frequent comparisons between lobsters and humans regarding power hierarchies.

Misconceptions About Biology

Peterson often draws parallels between lobsters and humans, suggesting that both species respond similarly to antidepressants by regulating neurotransmitters like dopamine and serotonin. However, these neurotransmitters serve different roles in lobsters compared to humans.

Nature uses neurotransmitters and hormones interchangeably for various functions across species. While closely related species may share similar functions for these chemicals, lobsters and humans diverged significantly long ago. As Leonor Gonçalves, a Research Associate in Neuroscience at UCL, explains:

"Our last common ancestor with the lobster existed 350 million years ago and was the first organism to develop an intestine—our main shared organ, not serotonin or the nervous system."

While lobsters might process neurotransmitters, it is a stretch to attribute emotions or feelings to them as we do with humans. Lobsters lack an amygdala, and their nervous systems do not resemble ours.

Jordan Peterson's Views on Climate Change

Peterson seems to relish making sweeping statements about topics he only superficially understands. At a 2018 Cambridge Union event, he positioned himself as an expert on climate change and energy policy. He stated that even if the most alarming claims about climate change were accurate, "we have no idea what to do about it."

He questioned the viability of transitioning to wind and solar energy, citing Germany's experience, claiming they produced more carbon dioxide after shifting to these renewable sources. However, this assertion is demonstrably false. Data from Our World in Data shows that Germany reduced its CO2 emissions by 40% between 1980 and 2020, despite population and GDP growth.

It's crucial to note that one major critique of Germany's energy policy relates to the shutdown of nuclear power plants during that same timeframe. This decision hindered their potential emissions reductions. Remarkably, they managed to decrease CO2 emissions while undertaking significant energy policy changes, unlike Denmark, which maintained its nuclear power infrastructure and showcased a dramatic reduction in fossil fuel use alongside increased wind energy adoption.

In essence, Peterson's assertion that no one has figured out how to reduce CO2 emissions is simply unfounded.

He also claims that climate models over the past 50 years feature such large error margins that we cannot gauge the impacts of our actions. However, research published in Science indicates that many models accurately predicted contemporary global surface temperatures, which have risen approximately 0.9°C since 1970. For ten forecasts, there was no statistically significant difference between model outputs and historical observations.

Climate models do not predict human behavior; instead, they offer projections based on various scenarios, allowing readers to grasp the severity of potential developments and necessary actions. Current models continue to forecast significant temperature increases under different human action scenarios.

The Flaws of Economic Models in Climate Policy

Peterson praises Bjørn Lomborg's climate work, despite Lomborg lacking credentials in climate science or economics. Lomborg, who holds a PhD in political science, writes on these topics and publishes without peer review. Critics, including the Union of Concerned Scientists, have heavily scrutinized his methodologies.

Lomborg's approach often involves comparing the costs of climate change mitigation against the societal costs of climate change itself. However, as noted by Undark magazine, economic models for climate change are fraught with issues. Unlike the relatively straightforward physical processes modeled in climate science, economic forecasting involves myriad unpredictable factors.

For instance, the Dynamic Integrated Climate-Economy (DICE) model, commonly used in this context, is simplistic enough to run on basic spreadsheet software. Its assumptions about the relationship between temperature rise and economic impact are overly simplistic, ignoring the non-linear nature of climate change's consequences.

The models often fail to account for potential catastrophes that exceed mere economic calculations. For example, a predicted 2-degree rise in temperature does not imply a uniformly elevated temperature across the board. Higher temperatures can lead to extreme fluctuations, risking the survival of entire ecosystems and human populations.

Conclusion: The Danger of Non-Expert Opinions

Jordan Peterson appears more interested in promoting his image as a contrarian than in grounding his claims in scientific truth. It is vital to approach the statements of well-known individuals with caution, especially when they engage in political agendas rather than sound scientific reasoning.

This caution should not diminish respect for intelligent people discussing fields outside their expertise. I frequently write about topics beyond my specialization, yet my lack of recognition limits my influence. If I held greater sway, I would hope people would critically evaluate my assertions.

What About Greta Thunberg?

Does this critique of non-experts extend to figures like Greta Thunberg? Not in this instance. Thunberg, unlike Peterson, clearly identifies as an activist rather than a scientist, emphasizing the importance of listening to experts in the field. Her advocacy aligns with a well-established scientific consensus, contrasting sharply with Peterson's extraordinary claims unsupported by robust evidence.

In summary, while challenging established views can lead to significant scientific advancements, those who do so must be well-versed in their fields. Thunberg’s approach reflects an awareness of her position, focusing on the scientific community's insights rather than claiming expertise she does not possess, unlike Peterson, who often presents himself as an authority on matters outside his field.

In the first video titled "New Atheists Are Not Complex | Jordan Peterson," Peterson discusses the misconceptions surrounding contemporary atheism and its implications for society.

The second video, "An Unfiltered Conversation with Jordan Peterson," features an in-depth discussion where Peterson shares his unfiltered views on various controversial topics.