Приближающийся AGI: Эксперты рассматривают потенциальные угрозы для рабочих мест Translation: Impending AGI: Experts Examine Potential Threats to Jobs

The emergence of Artificial General Intelligence (AGI) is approaching rapidly, leaving politicians with diminishing time to prepare. This was stated by Dario Amodei, the CEO of Anthropic, during a panel discussion at the World Economic Forum in Davos.

The entrepreneur cautioned that the swift advancement of AI may outpace the ability of labor markets and social institutions to adapt to these changes.

Amodei believes that transformations will occur within «a few years, not decades.» He reaffirmed a previously shared prediction that AGI could manifest by 2026 or 2027.

«I don’t think it will be a long time coming. It’s hard for me to imagine it taking longer,» Amodei emphasized.

The driving force behind the accelerated development of AI is self-learning, where models automate their own creation. The CEO noted that at his company, traditional programming roles are already being handled by artificial intelligence.

«I have developers who don’t write code. They let the neural network do it and then edit afterwards. In six months or a year, the LLM may handle most, if not all, of the work,» Amodei said.

In his view, progress is limited only by chip supply and training cycles.

Demis Hassabis, CEO of Google DeepMind, takes a more cautious stance.

«Significant progress has been made in certain areas. In fields like mathematics or programming, it’s easy to envision how AI can be automated since their results are verifiable. However, it’s more complex in the natural sciences. You can’t know if a chemical compound or a physics prediction is correct,» the expert remarked.

Hassabis added that current LLMs still struggle to generate original questions, theories, or hypotheses.

«This is the highest level of scientific creativity, and it’s unclear whether we will have such systems,» he noted.

The DeepMind leader estimates a 50% chance that AGI will appear by 2030.

Both executives reached a grim consensus regarding the economic implications. They believe that «white-collar» jobs are at risk.

According to Amodei, up to half of entry-level professional positions could vanish within a decade.

Hassabis warned that even the most pessimistic economists may underestimate the speed of transition.

«Five to ten years isn’t that long,» he reported.

Some analysts assert that changes will not come from direct job replacement but rather from a restructuring of professional activities.

«We need to stop asking whether AI will replace our jobs and start asking how it will degrade them,» said Bob Hutchins, CEO of Human Voice Media.

The impending shift changes the role of humans from «creators» to «verifiers,» the expert added.

«This strips professionals of the ability to make their own decisions and breaks meaningful professional work into unskilled, low-paying tasks,» he pointed out.

It’s worth noting that in August 2024, British trade unions warned of the risk of millions of job cuts due to artificial intelligence.