In his column for Fast Company, Yura Lazebnikov, Managing Partner of the international holding TECHIIA, explains why AI, in today's reality, is not a competitor to humans but rather a helpful assistant.
The boom of generative platforms has once again brought up the discussion of how the classifier of professions will change if artificial intelligence becomes smarter than humans. I’ll try to answer the question, “When will algorithms finally replace us and we won’t have to work?” from the perspective of an entrepreneur and tech optimist.
AI — a new horror story for adults
The idea of losing one’s job due to a new scientific and technological leap is both frightening and attractive. However, when reading news about AI achievements, few people imagine how they will sit back all day having handed over their tasks to intelligent machines.
I recall scenes from movies in which AI rises up against humans, such as in "2001: A Space Odyssey". In this film, HAL 9000 — a spacecraft control system based on artificial intelligence — kills an astronaut and shuts off oxygen to the researchers, who are in suspended animation.
AI creates a conflict in which humans are reduced to the role of a worn-out battery — they will be replaced by machines and discarded as scrap. Media fans the fires with headlines like "By 2033, Tesla's new microchip
ChatGPT is not ready to fire us
Personally, I don't mind if ChatGPT writes columns instead of me. But so far, the neural network is not able to handle this task. I asked it to write an essay on the topic "When will AI be able to completely replace all professions?" It wrote without errors but also without interesting examples, simply listing facts from other materials.
Most importantly, the text abruptly ended in mid-sentence. ChatGPT informed me that in one block, it can only publish 2048 characters. But when I asked it to remind me of the maximum message size, it gave me a different number — 2000.
— It was 2048!
— Sorry for the mistake. The maximum number of characters in one message is 2048.
This bot still has a long way to reach the level of HAL 9000. Having asked about the experience of other users, I
Neural network out-of-the-box is an assistant, not a competitor
I am confident that over time, ChatGPT will make fewer mistakes as its neural network constantly learns. However, today there are plenty of ready-made AI programs available that can be used not only for entertainment. For example, publishers working on releasing a new book in an audio format no longer have to conduct casting calls for voice actors. A speech-to-speech generator can create voices with the necessary tones. These same neural networks could be used to create voices for avatars when the concept of a metaverse takes off.
An IBM study
Again: these are not whitepapers, these are working neural networks that bring income to their creators. While you are reading, these projects are learning and processing new volumes of information, thus becoming even more functional.
Our holding also has an AI project, one of the main technologies of which can create digital models of people of any appearance. It has numerous applications, including e-commerce and the beauty sector. However, I still don’t see how neural networks will send humanity into retirement all at once. They can be assistants that optimize human work and free up time. They can add new professions and replace the most routine ones. But they cannot completely cancel them out.
AI will continue to develop, but it will not replace humans
So, we won’t be replaced even if all investments flow into the AI sector. Here are the most important obstacles.
1. A neural network cannot fully emulate the human brain. At least not yet. Computer vision and the ability to create predictive models do not make it a highly organized creature capable of thinking like a human.
2. Modern neural networks demonstrate an insufficient ability to learn on small amounts of data. Teaching AI on large amounts of data is too expensive (ChatGPT, by the way, learns for free, answering your questions). In addition, neural networks cannot interpolate the knowledge gained into new situations and tasks.
3. AI development can be hindered by regulators. The huge resources required to implement such projects, in my opinion, can make it difficult for new players to enter the market. This, in turn, can lead to the monopolization of the market by large players.
The latest high-profile cases in this field are the British
However, the AI market, including generative platforms, will continue to grow. In January, Microsoft
I consider hype, "borrowing" of venture investors from the crypto-currency segment, and requests from large companies to be the main drivers of the sector growth. However, the development of AI technology will happen gradually rather than in leaps and bounds. But we will all feel how our lives are changing. And no, the lack of jobs in the near future is not a threat to us.